Regression and capillary growth potential Regression of glomerulosclerosis has been achieved both in the ligation hypertensive remnant kidney model and in the nonhypertensive sclerosis associated with aging with high-dose angiotensin type 1 receptor blocker (ARB) or angiotensin I converting enzyme inhibitor (ACEI)[1]. Regression has also been achieved with combination ACEI, ARB and statin therapy in a severe passive Heymann’s nephritis model with sclerosis, and with ARB in the hypertensive nitric oxide-deficient rat model [2, 3]. Carefully detailed studies from the groups of Amann and Ritz have confirmed that regression can occur in the surgical cautery remnant kidney model in the rat with high-dose ACEI, with parallel regression of existing glomerular, tubular and vascular scarring, inferred by comparing sclerosis severity from sacrifice of different rats at varying time points after delayed intervention [4]. Additional studies by these investigators showed marked endothelial cell increase per glomerulus after injury, reversed by ACEI, with reduction of glomerular volume and capillary number [5]. Whether the same capillary remodeling mechanisms are involved in all models of sclerosis has not been established. Three-dimensional confocal images and theoretical mathematical modeling may become useful tools to analyze capillary branching in remodeling of sclerosis. So-called graph theory analysis of networks of capillary branching from serial section glass histologic slide data indicates that patterns of glomerular branching differ in uremic versus normal or newborn rat glomeruli [6]. Capillaries from uremic rat glomeruli showed a shift in vertex degree distribution, with more frequent higher vertices (ie more complex branching)[6].

How then can new capillary loops evolve from the segmentally sclerotic glomeruli, and are there limits to the potential of regression? Our previous data with mathematical modeling indicate a limit of the regression that can be achieved in individual glomeruli: if an individual tuft is sclerosed more than 50%, it is doomed to progression. Conversely, those glomeruli with less than 50% of the tuft sclerosed may grow new capillary loops [1]. Detailed data are not yet available on specific capillary branching patterns when regression is achieved. However, elegant morphometric studies in chronic kidney disease in children and in rats have demonstrated that both capillary lengthening and branching contribute to such glomerular growth after injury [7, 8]. Confocal microscopy technology has now advanced such that Z-sections from confocal images spanning an entire glomerulus are achievable, and could greatly enhance understanding of the complex glomerular structure. Proof of principle of regression has also been shown in human diabetic nephropathy, where cure of the underlying diabetes by pancreas transplant enabled regression of existing glomerulosclerosis and tubulointerstitial fibrosis over a 10-year follow-up period [9].