Abstract
Background. Haematuria of glomerular origin, even if mild, implies the development of defects in the glomerular basement membrane (GBM). In diseases where there is no infiltration of leukocytes into the glomerulus—such as thin basement membrane disease (TBMD) and histologically mild cases of IgA nephropathy (IgAN)—the mechanism by which such defects form is unclear. Methods. Frozen renal tissue from 18 cases of TBMD, 18 of mild IgAN and 18 cases with no detectable abnormality were studied: (i) by quantitative in situ zymography, to estimate the activity of glomerular collagenases; and (ii) by quantitative immunoelectron microscopy to estimate the amount of major basement membrane proteins per unit length and per unit area of glomerular basement membrane. Results. Cases of IgAN showed considerably more glomerular collagenase activity than normal (P=0.001). Thin basement membrane disease showed no difference in collagenase activity. A count of LCA‐positive cells in glomeruli confirmed that the IgAN cases did not show glomerular leukocyte infiltration. Conversely, cases of IgAN showed no difference in GBM composition from normal, nor was any difference in GBM thickness detected in this group. However, cases of TBMD showed considerably less laminin (P=0.0008), fibronectin (P=0.002) and type VI collagen (P=0.0005) per unit length of basement membrane. Collagen IV showed a smaller reduction per unit length (P=0.01), but unlike the other protein studies it appeared to be present in higher concentration per unit area (P=0.03), suggesting that it is more ‘compact’ in TBMD disease. Conclusions. Two distinct mechanisms of haematuria seem to be involved in these two conditions. In IgAN there is increased activity of enzymes that can degrade GBM, probably reflecting mesangial cell activation. In TBMD an abnormal composition of the thinned GBM is confirmed. When considered with published reports of genetic abnormalities in TBMD, these results raise the possibility of an abnormal interaction between collagen IV and laminin.