Glomerular permselectivity in early stages of overt diabetic nephropathy

Background. Impairment of glomerular size selectivity has been demonstrated by the dextran-sieving technique in nephropathic diabetics with heavy, but not mild proteinuria. The purpose of the present study was to determine whether such a barrier defect could be demonstrated with mild proteinuria by substituting Ficoll 70, a polysucrose, for dextran as a probe of the filtration barrier. Methods. Differential solute clearances were performed in 12 individuals with early diabetic nephropathy on two occasions: after 60 days of treatment with losartan 50 mg daily or a placebo. An uncharged preparation of nonreabsorbable Ficoll 70 was infused along with inulin. Fractional clearance (theta) of Ficoll of discrete size was determined after separating molecules in urine and plasma in narrow 2 Angstrom fractions over a 20 to 60 Angstrom radius interval by size exclusion high-performance liquid chromatography (HPLC). A hydrodynamic theory of hindered ficoll transport through water-filled pores was used to characterize the pore size distribution of the glomerular barrier. Results. The theta for Ficoll molecules with radii >50 Angstrom was selectively enhanced in placebo-treated diabetic nephropathy versus corresponding theta in healthy control subjects (N = 12). Computations revealed a lower distribution of glomerular pores that was unaltered in nephropathic diabetics. However, an upper distribution of large, shunt-like pores was more prominent, exceeding healthy controls by one order of magnitude in diabetic nephropathy (P = 0.01). A trend to lower theta for Ficoll in the 56 to 60 Angstrom radius range during losartan therapy is computed to have lowered the fraction of shunted filtrate by 26 to 44%, depending on whether glomerular pressure declined. The corresponding reduction in theta for endogenous albumin, IgG, and IgG(4) was by 19 to 23% (P < 0.05). Conclusion. Our findings suggest that shunt-like defects, partially reversible by angiotensin II blockade, are present early in the course of diabetic nephropathy. We estimate that such defects can account for immunoglobulinuria in this disorder. Additional impairment of either charge- or shape-selectivity must be invoked to explain the observed level of albuminuria, however.