Perspectives in Basic Science

PHYSIOLOGIC MECHANISMS OF GLOMERULAR FILTRATION AND TUBULAR REABSORPTION OF PROTEINS

Determinants of the transglomerular passage of plasma proteins

Despite the extremely low resistance to water flow, the glomerular capillary wall very efficiently restricts the passage of proteins from the blood into Bowman's space on the basis of their molecular size, electrical charge, and sterical configuration. Large and negatively charged molecules are less readily filtered than small and electroneutral or positively charged ones. The mechanisms whereby normal glomerular capillary wall restricts the transmural passage of large plasma proteins have been extensively explored in the last 25 years, but a universally accepted theory has not yet been agreed upon. There is a consensus that the glomerular wall may be thought of as containing three barriers in series. Along its route to this filtration barrier, the filtrate first has to pass through the open pores of the glomerular endothelium, then through the highly hydrated collagenous network of the glomerular basement membrane (a molecular scaffold composed of tightly cross-linked type IV collagen, laminin, nidogen, and proteoglycans), and finally, through the filtration slits established by interdigitations of podocyte foot processes and covered by a complex structure called "slit diaphragm," which is anchored at the sides of the adjacent foot processes^1,2,3,4,5,6. Although the fenestrated endothelium can represent an electrostatic barrier for negatively charged proteins, and glomerular basement membrane with its laminae can restrict the traversal of large plasma proteins and negatively-charged proteins (due to the presence of heparan sulfate proteoglycans), recent evidence suggests that the ultimate, more selective barrier for the majority of proteins resides in the slit diaphram^7,8,9. However, the molecular composition of this last structure is still largely unknown. The most recent studies tend to confirm the model of substructure proposed by Rodewald and Karnovsky¹⁰ in 1974 on the basis of their transmission electron microscopic findings, according to which the slit diaphram is made up of rod-like units connected in the center to a linear bar, forming a zipper-like pattern. Nephrin, a molecule specifically expressed in podocytes, has been suggested by Tryggvason⁸ to be the candidate molecule that might assemble into a zipper-like isoporous filter structure. Nephrin molecules extending toward each other from two adjacent foot processes are likely to interact in the slit through homophilic interactions and covalent cross-linking. Other proteins, such as ZO-1¹¹, P-cadherin, and catenins⁹, have also been found to be expressed in the slit diaphragm and could be involved in the structure of the adherent junctions of a zipper-like filter⁹. Very recently, two other proteins have been discovered as components of the slit diaphragm domain of podocytes, CD2AP, which for its structure probably acts as an adaptor linking the actin-based cytoskeleton of podocyte to nephrin¹², and podocin, a stomatin-like protein that forms aggregates and organizes lipid rafts¹³. A complex of nephrin, podocin, and CD2AP is emerging as the functional unit upon which the slit diaphragm assembles. These proteins are tightly associated and are embedded into lipid rafts. Podocin may be critical for the stability of this complex; its binding dramatically activates the signaling capabilities of nephrin^13,14,15,16. The basal portion of podocytes is attached to the glomerular basement membrane through several adhesion proteins. The integrin ₃₁ and the dystoglycan complex are the major responsable for the binding to the extracellular matrix of such basement membrane^14,15,16. ₃₁ integrin constitutes stable, static bonds, and is associated on its cytoplasmic side with paxillin, talin, and vinculin, which mediate its connections to the actin cytoskeleton. The dystoglycan complex probably provides an actin-directed positioning system by which podocytes activity control the exact spacing of matrix proteins, and thus the porosity and permeability of the glomerular basement membrane¹⁵.

In the last 20 years, numerous studies have been performed, in experimental animals and in humans, based on the measurement of the fractional excretion (sieving coefficient) of some biologically inert molecules of different size, either neutral or charged, used as test transport probes. They demonstrated that glomerular capillary wall, as a whole, can be assumed to be perforated by cylindrical pores of different diameter. The most used have been for years the dextrans of different size. They have given useful results for the definition of the porosity of the glomerular filter. However, like many of the proteins physiologically crossing this filter, they are deformable and flexible (differences in molecular configuration may have an important influence on the relative filtration rates of proteins); therefore, they did not allow to precisely define the real radius of the pores. More recently, another inert molecule, Ficoll, has been preferred for such studies, because it is a rigid spherical molecule, the configuration of which is not altered during transglomerular passage. Many different mathematical models, based on hemodynamic theories, have been proposed in order to explain the results of the studies on the transport probes. The models most extensively accepted, called heteroporous models, consider the glomerular capillary wall as a membrane perforated by pores of different diameter having a log-normal distribution of radii^{17,18,19,20,21,22,23,24,25,26,27,28,29}. A single population of restrictive pores with a log-normal distribution of radii extending to radii>60 ÅFigure 1a has been recently considered the best fitting model by Ruggenenti et al²⁸. In contrast, others^18,20,24,27, using as transport probe not only dextrans, but also Ficoll, have recently concluded that a bimodal model better fits the description of their results Figure 1b. They hypothesized a prevailing population of small restrictive pores with a log-normal distribution of radii, with a mean radius around 45 Å, with a range between 37 and 48 Å, and a limited number of nonrestrictive pores (>80 Å), that they called the "shunt" pathway, that in physiologic conditions is permeated by a very small fraction of filtrate (only 0.1 10^-3 on average). However, even among the supporters of the bimodal distribution, not everyone accepts the nonrestrictive nature of the large shunt-like pores. A group of investigators from Lund^25,26,29, using a different approach to the study of glomerular permeability based on the inhibition of tubular reabsorption in animal models, agrees on the existence of two populations of pores, the vast majority being "small pores," but hypothesizes that even the second pore population consists of a limited number of discriminatory "large pores" with a radius 80 to 90 Å. These investigators postulate that, in addition to the two populations of pores, sporadic "membrane defects" or "shunts" large enough to allow transport of very large proteins of the size of ₂-macroglobulin and even red blood cells (similar to the non-discriminating shunt-like pores proposed by the Myers group) could also exist, and that they should contribute less than 10^-5 of the total glomerular filtration rate (GFR) in physiologic conditions, but could increase in number in pathologic conditions ("two pores with a shunt" model).

Figure 1.

Glomerular pore size distribution in normal individuals (solid curves) and in nephrotic patients (dashed curves), according to the two most accepted theoretical models. (A) The membrane is perforated by a single population of cylindrical restrictive pores that have a continuous log-normal distribution of radii (heteroporous model). (B) The membrane is perforated by a lower distribution of restrictive pores with a log-normal distribution of radii and a parallel upper distribution of nonrestrictive dimensionless shunt-like pores (bimodal heteroporous + shunt model).

Full figure and legend (26K)

It is accepted by all investigators that in physiologic conditions, proteins of the size of immunoglobulin G (IgG) (molecular radius = 55 Å) are completely (or almost completely) restricted from filtration because their radius is higher than that of all the small restrictive pores, and the contribution of the shunt pores or large selective pores is quantitatively irrelevant Figure 2a. On the contrary, the low permeability of the glomerular capillary wall to albumin (molecular radius = 36 Å) cannot be explained in terms of simple restriction by pore size. The majority of investigators refer the restricted movement of this protein to the Bowman's space to some characteristics related to its negative charge and to the consequent electrostatic repulsive interactions with the negative charge of the glomerular capillary wall. It has been demonstrated experimentally that some transport probes, if negatively charged, cross the glomerular filter with a reduced sieving coefficient and that neutralization of the anionic sites of glomerular basement membrane by polycations results in increased permeability to such probes, as well as to albumin. It has been recently calculated, using uncharged Ficoll as a probe, that, in rats, albumin has a much lower sieving coefficient compared with Ficoll of the equivalent size of 36 Ų⁸ and that in normal individuals albumin filtration is restricted by a factor of approximately 100 relative to a 36 Å probe²⁴. However, the role of charge selectivity in permselectivity of glomerular capillary wall has been recently questioned²². The anionic charge of the capillary wall has been estimated to be much lower than that required to affect the clearance of some proteins based on charge repulsion. Even the higher fractional clearance of cationized dextran in comparison with its anionic counterpart found in many in vivo experiments has been explained arguing that this cationic polysaccharide binds to anionic sites and damages the glomerular capillary wall, making it leaking²².

Figure 2.

Schematic representation of transglomerular transfer to the tubular lumen, reabsorption by the tubular cells and excretion in the urines of plasma proteins of low-molecular-weight (LMW), intermediate-molecular-weight [mainly albumin ALB)], and high-molecular-weight (HMW) in physiologic conditions and in pathologic conditions characterized by progressively increasing permeability of the glomerular barrier and reabsorptive load of the tubular cells. (A) In physiologic conditions, all LMW proteins and a fraction of albumin cross the glomerular barrier and are completely reabsorbed by the tubular cells. (B) The alteration of the permeability of the glomerular barrier is moderate, involving mainly a loss of restriction to passage of negatively charged proteins (especially albumin); albumin and a small fraction of HMW proteins reach the tubular lumen and saturate the reabsorptive capacity of tubular cells, inducing the loss in the urines of a fraction of LMW proteins and albumin, together with a very small fraction of HMW proteins (selective proteinuria).

(C) A more severe damage progressively increases size permeability of the glomerular barrier, and, due to the saturation of the reabsorptive mechanisms of tubular cells, a greater percentage of HMW proteins is excreted in the urines (nonselective proteinuria). (D) Permeability of the glomerular barrier is further increased, and the massive and protracted reabsorptive load of the tubular cells induces toxic lesions of these cells and reduces their reabsorptive capacity; excretion with the urines of all three classes of proteins and, in particular, of LMW and HMW proteins, is increased, and represents a valid marker of the severity of the glomerular and tubular damage.

Full figure and legend (69K)

Renal tubular handling of filtered proteins

As we have stated, in normal conditions only a fraction of proteins of intermediate molecular weight, in particular albumin, the concentration of which in the Bowman's space of healthy animals has been calculated to be approximately 1 mg/dL^30,31 or even higher³², and almost none of the high-molecular-weight (HMW) protiens arrives in the tubular lumen. On the contrary, the passage across the glomerular filter of all proteins of molecular weight lower than 40,000 D and radius lower than 30 Å, the so-called low-molecular-weight (LMW) proteins, is almost completely unrestricted in the normal individuals. All these proteins that arrive in the tubular lumen are excreted in negligible amounts in the urines in physiologic conditions because of a very efficient mechanism of total reabsorption by the epithelial cells of proximal tubules Figure 2a. For ₂-microglobulin, the molecular radius of which is 11.8 Å, the block of tubular reabsorption with lysine in normal rats increased the clearance by a factor of 450, while the clearance of large proteins increased only five- to sixfold²⁰.

This reabsorption of proteins occurs predominantly in the pars convoluta (S1 and S2 segments) and, to a lesser extent, in the pars recta (late S2 and S3 segments) of the proximal tubule. The epithelial cells of these segments contain an extensive apical endocytic apparatus, consisting of a network of coated pits and small coated and noncoated endosomes, but also of prelysosomes, lysosomes, and so-called dense apical tubules involved in recycling membrane from the endosomes to the apical plasma membrane. The proteins absorbed at the luminal membrane are endocytosed and concentrated within vesicles at the apical border of tubular cells (endocytic vesicles). These vesicles fuse with acidic organelles that belong to the endosomal compartment. The endosomes containing the segregated proteins then migrate to the cell interior, where they fuse with the lysosomes, cell organelles that contain acid hydrolases. Absorbed proteins are completely hydrolyzed within lysosomes and the resulting aminoacids cross the contraluminal membrane to return to the circulation^33,34.

Until recently, it was believed that proteins were largely absorbed by nonspecific absorbptive endocytosis, the tubular uptake being a high-capacity, low-affinity transport system. However, kinetic parameters differ significantly according to the charge of proteins, and, even for proteins of similar size and net charge, there is competition for tubular absorption among the various proteins. Moreover, in the case of albumin, in addition to the high-capacity, low-affinity transport system, which is universal among all proteins tested, a low-capacity transport system has been found that is probably operational in the lower ranges of tubular fluid albumin concentration^35,36,37. There is a growing body of evidence suggesting that the absorption of albumin, as well as all other proteins, through the entire length of the proximal tubule but especially in its pars convoluta, is "receptor-mediated" as opposed to "nonspecific." It is assumed that the luminal endocytosis is initiated by ligand binding to receptors localized in the clathrin-coated pits, followed by internalization, segregation of ligands and receptors in early and late endosomes, and directing of ligands to lysosomes for degradation, while the receptors are directed back to the apical plasma membrane via dense apical tubules³⁴. Data gathered in recent years indicate that megalin and cubilin play a crucial role in this process³⁸. Both are multiligand receptors massively expressed in the proximal tubules (but also in several other epithelia). Although they belong to distinct protein families, they share a common structural feature, the accumulation of motifs potentially important for ligand binding, which confers upon them a key property for multiligand receptors. Both are detectable over the apical endocytic apparatus: clathrin-coated pits, small and large endosomes, and dense apical tubules. Megalin and cubilin appear to carry on a cooperative function, clearly demonstrated in the case of the absorption of albumin, where this apparently redundant interaction results in a very efficient uptake, even though in the case of transferrin, cubilin appears to be exclusively involved in the internalization process. Megalin and cubilin, although putative receptors, do not appear to be very selective, as many proteins are able to bind to them. Because of this multiplicity of ligands, it would probably be more appropriate to characterize them as "binding proteins" rather than as "receptors"³⁴. A potential role of megalin to signal into the cell interior to stimulate cell proliferation has been very recently hypothesized³⁹.

It is worth stressing that cationic proteins, such as IgG, have been shown to bind more avidly to the apical membrane of proximal tubule cells because such membrane is negatively charged. As a consequence, cationic proteins present in the tubular lumen are more readily endocytosed than anionic proteins such as albumin^40,41. This competitive mechanism adds another variable that affects the concentration of filtered proteins in the final urine, and renders the correct calculation of the relative sieving coefficient of HMW and LMW proteins and of the selectivity index more difficult, especially in pathologic conditions.

PATHOPHYSIOLOGY OF ABNORMAL URINARY PROTEIN EXCRETION IN GLOMERULAR DISEASES

Increases in the urinary excretion of proteins result from increases in their filtered load, due to alterations in the permselectivity of the glomerular capillary wall, or from defects in their tubular uptake. The majority of the experimental studies using inert test probes, reported above, showed that the alteration of the selectivity of the glomerular capillary wall was a combination of the loss of charge restriction and size restriction. These studies did not necessarily show an increase in the mean radius of the small restrictive pore population, but rather, in all experiments, demonstrated a spread of their distribution due to the appearance of more and wider large-pore pathways (either discriminatory or shunt-like pores) Figure 1. According to Tencer et al⁴², even for the large pores, a charge restriction exists in physiologic conditions, and explains why, especially in membranous nephropathy, the ratio between urinary concentration of neutral IgG2 and negatively charged IgG4 is significantly reduced, indicating a loss of the charge restriction for IgG4.

The reduction in the restrictive properties of the glomerular barrier leads to proportionally greater increases in the filtered loads of albumin and HMW proteins than of LWM proteins; the glomerular permeability of the latter is already very high and cannot increase further, whereas even a small increase in glomerular permeability to macromolecules leads to very significant increases in the filtered load of larger proteins.

Increased transglomerular passage and urinary excretion of intermediate and HMW proteins

A moderate increase in the permeability of the glomerular capillary wall is characteristically found in minimal change nephropathy and in some initial stages of other glomerular diseases (primary focal segmental glomerulosclerosis, membranous nephropathy, diabetic nephropathy). The increased transglomerular passage in the tubular lumen of intermediate-molecular-weight proteins, mainly albumin, is not accompained by a similar passage of HMW proteins, such as IgG, ₂-macroglobulin or IgM. Despite their partial reabsorption by the tubular cells, a fraction of albumin and intermediate-molecular-weight proteins escapes the reabsorptive process and appears in the urine. The ensuing proteinuria is indicated as "selective"Figure 2b. A selectivity index (SI) based on the comparison of the clearance of IgG, as a marker of proteins of HMW, and that of transferrin, as a marker of proteins of intermediate size, had already been elaborated in 1964, and since then patients with an IgG SI of 0.2 or higher have been considered to have a nonselective proteinuria, whereas patients with a ratio below 0.2 have been considered to have a selective proteinuria⁴³. In spite of the criticism to the theory of charge restriction, to which we alluded in a previous section, the majority of investigators still believe that in these pathologic conditions characterized by selective proteinuria, the impairment of the charge selectivity of the glomerular filter is prevalent in comparison with the impairment of size. The injury to the structures of the glomerular capillary wall impairs their function as a negatively charged electrostatic barrier, permitting the unrestricted transglomerular passage of albumin, while the damage responsable for the increase of the large nonselective pores permeating the capillary wall is less marked. A loss of anionic sites of the glomerular basement membrane has been documented in minimal change nephropathy^44,45 and in the microalbuminuric stage of diabetic nephropathy⁴⁶. In both conditions the study with dextrans has confirmed a very mild increase of the area of the glomerular filter being permeated by large pores, while the transglomerular passage of albumin through the small selective pores is more markedly increased^18,47,48,49.

In all the glomerular diseases with nonselective proteinuria, the variable quantity of proteins of HMW reaching the tubular lumen is an expression of a variable severity of the impairment of the size selectivity, in addition to that of the charge selectivity Figure 2c and d. The sieving coefficient of large (>60 Å) dextrans has been found to be more increased in patients with focal segmental glomerulosclerosis, in comparison with patients with minimal-change nephropathy, in the study already mentioned¹⁸. Another study of the same group⁵⁰, comparing the sieving coefficients of uncharged dextran and anionic dextran sulfate in nonproteinuric controls and in two groups of nephrotic patients, with minimal-change nephropathy and membranous nephropathy, confirmed that the permeability to large uncharged dextrans, present in all nephrotic patients, was more evident in patients with membranous nephropathy. The sieving coefficient of the anionic dextran sulfate of small radius (18 Å) was definitely lower than that of uncharged dextrans of similar radius in nonproteinuric controls, but not in nephrotic patients, indicating a loss of charge restriction which could explain the magnitude of albuminuria better than the concomitant increase in filtration of large neutral dextrans.

More recent studies of the same investigators from Stanford, using Ficoll test probe, have been performed in membranous nephropathy²⁴ and IgA nephropathy⁵¹. In comparison with dextrans, sieving coefficients for all restricted Ficoll molecules (probably because their conformation was not altered during transglomerular permeation) were uniformly lower than those of dextrans in nonproteinuric controls²⁴. In patients with membranous nephropathy, the membrane parameters measured with Ficoll differed from control in two respects: (1) the distribution of restrictive pores was broadened, the mean radius (36 Å) being shifted to smaller size (from 44 Å to 36 Å); (2) the magnitude of the area being permeated by large shunt-like pores was substantially increased, the sieving coefficient for Ficoll of equivalent radius to IgG (54-56 Å) being enhanced above control by a factor of 20. Since the sieving coefficient of Ficoll of equivalent radius to albumin (36 Å) did not differ between the patients with membranous nephropathy and control groups, the authors suggested that "massive immunoglobinuria is attributable to an impairment of barrier size selectivity, whereas the massive albuminuria is due to a defect of charge selectivity"²⁴. In patients with IgA nephropathy⁵¹, whose proteinuria was less severe than that of patients with membranous nephropathy, the study with Ficoll confirmed that the fraction of filtrate permeating the large shunt-like pores was significantly increased in comparison with non-proteinuric controls. At variance with the results obtained in nephrotic patients with membranous nephropathy, no differences from control subjects were found in the mean radius of the small restrictive pores or the range of distribution of these pores. There was a significant relationship between the fractional clearance of albumin and IgG and the fraction of filtrate passing through large shunt-like pores in this population of patients⁵¹.

Similar studies^{28,29,47,48,49,52,53} using neutral dextrans or Ficoll as probes have been carried out in the various stages of type 1 or type 2 diabetic nephropathy. These studies showed that the evolution of the disease, through the microalbuminuric stage, to one of overt proteinuria, represents a continuum of progressively deranged glomerular capillary wall function. As we said, in the microalbuminuric stage all parameters of size selectivity (mean radius and spread of radius distribution of restrictive pores, as well as the magnitude of the large pores) were unaltered or moderately altered. With increasing severity of the nephropathy, in the stage of overt proteinuria, the number and peak radius of the restrictive small pores was still unaltered or moderately increased. The increased filtration of IgG and other proteins of HMW could be explained with an increased fraction of filtrate being permeated by large pores. In type 2 diabetes of Pima Indians, it was calculated, using dextrans as probes, that an excess of large pores becomes evident when albuminuria reaches a level of 3 grams per gram of creatinine⁴⁹. In the same disease in Italian patients, using the same probes, Ruggenenti et al²⁸ concluded that distribution of pore sizes was on average shifted toward larger pores, as compared to normal controls, the mean pore size radius being of 5 Å greater Figure 1a. The majority of investigators agreed that, when overt proteinuria develops in the nephropathy of both types of diabetes, concomitant changes of charge—and size—selectivity are likely to contribute to albuminuria. However, Bakoush et al²⁹ found a better preserved ratio of IgG2/IgG4 urinary excretion in European patients with type 2 diabetes than in those with type 1 diabetes and argued that in the former group charge selectivity is less impaired even in the macroalbuminuric stage. They suggested that in type 2 diabetes even albumin is lost primarily through the large pore pathway. According to these investigators, even urinary excretion of IgM, a protein of molecular weight and radius higher than IgG, was more frequently detectable in type 2 than in type 1 diabetes; this difference was explained as reflecting a more marked increase in the population of the nonselective shunt-like pores in type 2 diabetes.

Structural-functional correlations have been analyzed in both diabetic and nondiabetic nephropathies associated with documented changes in size selectivity.

In overt diabetic nephropathy of Pima Indians, Lemley et al⁴⁹ documented increased glomerular basement membrane thickness, greater foot process width, and a decreased number of visceral epithelial cells per glomerulus; at multivariate regression analysis, only foot process width was significantly correlated with the increase in larger size pores.

In nondiabetic glomerular diseases characterized by prevalent lesions of the glomerular capillary wall, like membranous nephropathy and focal segmental glomerulosclerosis, the changes in charge and size selectivity have been ascribed to alterations of foot processes and slit membranes of epithelial cells, especially when effacement of adjacent foot processes or complete detachment of some foot processes and their slit membranes from the lamina rara externa of the basement membrane are detected^27,54,55, but also to the alterations of the structural modeling of the glomerular basement membrane^56,57. We have already mentioned the loss of anionic sites from such basement membrane and from podocytes, responsible for charge selectivity in some proteinuric diseases. Decreased density of slit processes of glomerular epithelial cells has been also documented morphometrically in human disease with nephrotic syndrome⁵², probably favored by the limited capacity of mature glomerular epithelial cells to proliferate⁵⁸. Dedifferentation of glomerular epithelial cells in some proteinuric diseases, especially focal segmental glomerulosclerosis, is suggested by the decreased expression of surface antigens usually present in mature cells^59,60. In glomerular diseases with mesangial involvement, such as IgA nephropathy, data are more controversial. Ikegaya et al⁶¹ found that the increase in pore size correlated with mesangial sclerosis and tubulointerstitial damage, more than with the severity of capillary changes. On the contrary, in their very accurate study, already quoted, Lemley et al⁵¹ have found that a reduced number of podocytes/glomerulus was the most significant morphometric feature associated with other markers of increasing disease severity such as loss of size selectivity of the glomerular barrier, IgG urinary fractional excretion, extent of glomerular sclerosis and decrease of GFR; they suggested that, whereas IgA nephropathy is initiated by intramesangial deposition of IgA, it is the injury to, and loss of, podocytes from the outer aspect of the glomerular basement membrane that determines the severity of disease as well as its rate of progression. In a model of lupus nephritis, the fractional clearance of albumin and IgG was significantly correlated with the percentage of glomerular basement membrane occupied by dense deposits, by the slit number/ glomerular basement membrane surface and by the thickness of glomerular basement membrane⁶². Finally, in all human glomerular diseases, when the injury to the glomerular capillary wall was severe, morphologic defects of glomerular basement membrane have been described, such as tunnels^63,64 or focal areas of rupture or holes in the wall^65,66. They are probably responsible for the leakage of very large proteins ("shunt pathway") and might allow the passage of blood cells.

It must be stressed that not only structural lesions, but also glomerular hemodynamic changes, affect the permeability of glomerular capillary wall to macromolecules. Filtration of proteins is influenced not only by the intrinsic membrane properties of the glomerular capillary wall, but also by the other determinants of single-nephron GFR: the glomerular capillary plasma flow rate (Q_A), the glomerular transcapillary hydraulic pressure difference (P), and the afferent arteriolar plasma protein concentration (C_A). Moreover, it is becoming increasingly apparent that hemodynamic alterations, either of pressure or flow, may directly influence the permeability properties of the barrier itself, as reflected by changes in the apparent pore size distribution and/or the concentration of fixed negative charges on the membrane⁶⁷.

The studies in normal rats^67,68,69, even after infusion of angiotensin II^{70,71,72,73,74,75}, or stimulation of endogenous angiotensin by renal vein constriction⁷⁶, have permitted better characterization of the effect of hemodynamic changes. It has been found that (1) selective increase in glomerular plasma flow rate (Q_A), up to a certain level, tends to decrease the fractional clearance of macromolecules (an effect due to a disproportionately greater increase in the average transcapillary volume flux than solute flux), while a decline in Q_A has the opposite effect; (2) selective increase of transcapillary hydraulic pressure gradients (P) leads to a decrease of the fractional clearance of macromolecules, even through it induces an increase in diffusion-driven macromolecule flux because it leads to an even greater increase in water flux; however, when P increases markedly, in response to a decline in plasma flow rate (Q_A), to maintain a normal single-nephron GFR, the consequent increase in filtration fraction induces an increase in the fractional clearance of proteins; (3) intravenous infusion of angiotensin II leads to proteinuria, concomitant and partially related to hemodynamic changes characterized by a significant decrease in Q_A and by a rise in P, with significant rise in filtration fraction; the degree of proteinuria is strictly correlated with the changes of Q_A and filtration fraction; (4) stimulation of intrinsic angiotensin with renal vein constriction produces similar hemodynamic changes and proteinuria, due mainly to increased efferent arteriolar resistance, P, and filtration fraction; (5) infusion of angiotensin II or stimulation of intrinsic angiotensin II markedly increase the fraction of filtrate passing through larger, nonselective pores; thus, changes in P play an important role in glomerular permselectivity since on increase in P recruits previously unexposed large pores of the nonselective shunt pathway; and (6) saralasin, an inhibitor of angiotensin, counteracted the effects of angiotensin II stimulation by renal vein constriction, induced reduction of efferent arteriolar resistance and P almost to preconstriction levels, normalized filtration fraction, decreased proteinuria, and corrected the barrier size permeability changes described above⁷⁶.

When angiotensin-converting enzyme (ACE) inhibitors, and, more recently, angiotensin I (AT₁) receptor antagonists became available as antihypertensive agents, and their antiproteinuric properties were confirmed in many experimental models of renal disease^{77,78,79,80,81,82,83} and in all human glomerular diseases^{52,53,58,84,85,86,87,88,89,90,91}, it was initially hypothesized that their effect on glomerular capillary wall permeability was due to the correction of the arterial hypertension and of the intraglomerular hemodynamic changes induced by the activation of the renin-angiotensin system (RAS) that we have summarized. However, correction of arterial hypertension using any other conventional antihypertensive drug does not influence the extent of proteinuria, the only exception being the nondihydropiridine calcium channel blockers, verapamil and diltiazem^92,93,94. Moreover, it is becoming evident that even the correction of the intrarenal hemodynamic alterations existing in proteinuric renal diseases does not fully explain the antiproteinuric effect of angiotensin inhibitors and nondihydropiridine calcium channel blockers. ACE inhibitors or AT₁ receptor antagonists exert their effect in animal models of nephritis despite unchanged intraglomerular pressure⁷⁹, or in experimental nephritis characterized by a normal capillary hydraulic pressure⁸⁰. The long-term effect of ACE inhibition on proteinuria in human glomerulonephritis could not be acutely reversed by intravenous administration of angiotenin II, despite a dose-related fall in renal plasma flow and rise in systemic blood pressure, renal vascular resistance, and filtration fraction⁹¹. An AT₁ receptor antagonist could prevent the alteration in permselectivity of the glomerular barrier induced by exogenous angiotensin II even in an isolated perfused rat kidney preparation in which hemodynamic modification had been minimized⁹⁵ As for calcium channel blockers, it is well established that they do not have a stable effect on efferent arteriolar tone and induce only little change in intraglomerular pressure and filtration fraction^93,96,97.

ACE inhibitors and AT₁ receptor antagonists seem to act directly on the permselectivity properties of the glomerular barrier. These drugs not only reduce the total amount of proteinuria, but can partially correct the increase in the fraction of filtrate being permeated by large shunt-like pores^{52,53,58,80,82,83,88,90,95}, and therefore reduce the fractional clearance of HMW proteins, such as IgG. The same reduction in the clearance of large dextrans, with parallel reduction of the fractional clearance of IgG, has been found recently in diabetic nephropathy after 6 months to 1 year of administration of the nondihydropiridine calcium channel blocker, diltiazem, but not of nifedipine. Only diltiazem induced a significant immediate reduction of proteinuria that could not be explained as due to changes in intrarenal hemodynamic different from those induced by nifedipine⁹⁸.

In the case of RAS inhibitors in general, an effect on partially reversible alterations in the structure and function of the glomerular capillary wall and of mesangial cells and matrix, induced by intrarenal angiotensin II, has been hypothesized: rearrangement of junctional complex of podocyte filtration slits, due to actin polymerization in the cytoskeleton of podocytes^95,99; reduction of the increase in the level of glomerular protein and cortical mRNA for transforming growth factor- (TGF-), and collagen types I and III⁸¹; and blunting of the marked increase in cell size of cultured cells exposed to angiotensin II¹⁰⁰. The fact that the antiproteinuric effect of angiotensin inhibitors may persist for some time after withdrawal of treatment⁹⁰ speaks in favor of a slowly developing improvement of structural alterations. In the case of nondihydropiridine calcium channel blockers, an effect on the metabolism of mesangial cells^101,102 and, in diabetic animals, an attenuated decrease in both glucosaminoglycan and heparan sulfate of the glomerular capillary wall and mesangial cells¹⁰³, have been described.

Renal tubular handling of an abnormal load of filtered proteins

In physiologic conditions, HMW proteins are restricted almost completely from reaching the tubular lumen because of the impermeability of the glomerular barrier Figure 2a. More controversial is the behavior of albumin. Everyone agrees that part of it crosses the barrier and arrives to the Bowman's space, where its concentration is at least 1 mg/dL^30,31; however, recent studies suggest that the filtration of albumin is very partially restricted by charge and its concentration in the tubular lumen is even higher³². Finally, it is universally accepted that the passage in the tubular lumen of the proteins of molecular weight lower than 40 kD and radius lower than 30 Å, the so-called LMW proteins, is completely unrestricted in physiologic conditions.

When impairment of charge and size selectivity of the glomerular capillary wall increases the filtration of the proteins of intermediate and HMW, these proteins compete among themselves, and with the LMW proteins, in the process of reabsorption by the epithelial cells of the proximal tubule. If this mechanism is saturated, some of the filtered proteins of all sizes, including those of LMW, appear in the urine Figure 2b to d. Unfortunately, the fraction of the different proteins that escapes reabsorption is not proportional to the filtered fraction, since it is becoming evident that the reabsorption process is not aspecific and unselective, and that even the charge of the various proteins influences such processes, IgG being more easily reabsorbed than albumin because of its positive charge.

When the load of filtered proteins escaping the glomerular barrier is massive and/or protracted, due to chronic glomerular injury, the epithelial cells of the proximal tubule, subjected to a continuous overload, may progressively lose their integrity, with impairment of lysosomal function and morphologic changes (enlargement of the protein absorption droplets and loss of brush-border structure). As a result, the reabsorption of all proteins, including the physiologic reabsorption of LMW proteins, is increasingly impaired, leading to the urinary excretion of a progressively larger fraction of LMW proteins Figure 2c and d. It has been found that such impairment of the reabsorption of LMW proteins, in glomerular diseases with severe and/or protracted proteinuria, correlates with the extent of the injury of the tubular cells and with the overall tubulointerstitial damage^{104,105,106,107,108,109}. Accumulating evidence suggests that direct toxicity for the tubular cells of some as yet nonidentified proteins reaching the tubular lumen might be responsible for the injury of such cells, triggering the release of a number of proinflammatory cytokines and growth factors, which cause an inflammatory and fibrotic response in the interstitium^{108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124}.

QUANTITY AND QUALITY OF PROTEINURIA PREDICT THE CLINICAL COURSE AND THE RESPONSE TO THERAPY IN GLOMERULAR DISEASES

Predictive value of amount of proteinuria and of tubulointerstitial lesions

Data accumulated in the last decade have demonstrated that in all glomerular diseases the amount of proteinuria, together with the degree of the tubulointerstitial damage, is the most powerful predictor of the progression to chronic renal failure^{107,109,110,125,126,127,128,129}.

The analysis of the pathophysiologic mechanisms of proteinuria performed in the preceding section gives a good theoretical explanation of the value of these two parameters as prognostic indicators. The extent of proteinuria is an indicator of the severity of the alterations of the glomerular capillary wall and of its permeability, and, therefore, can be a good marker of the overall severity of the glomerular damage. The consequent involvement of the tubular cells in the reabsorptive mechanisms of the proteins reaching the tubular lumen can explain the injury of these cells and, therefore, the correlation of the overall tubulointerstitial damage with the severity and duration of proteinuria.

These pathophysiologic mechanisms can also explain the recent finding that some qualitative characteristics of proteinuria are better predictors of the progression to chronic renal failure than its quantity.

Predictive value of selectivity of proteinuria and urinary excretion of HMW proteins

The assessment of SI, based on the comparison of the clearance of IgG and transferrin, has found little place in clinical practice for many decades, since it did not appear to help in predicting histologic diagnosis or in determining prognosis, the only exception being the differential diagnosis between minimal change nephropathy, in which proteinuria is always selective, and focal segmental glomerulosclerosis, in which proteinuria is often nonselective. However, more recently, Mallick, Short and Manos¹³⁰ showed that, even in membranous nephropathy, the selectivity index could discriminate between patients who remitted and those whose proteinuria persisted, the former having a SI lower than 0.19 Table 1. Woo et al¹³¹ confirmed the predictive value of SI in 98 patients with IgA nephropathy Table 2. We¹³² calculated the SI in 89 patients with nephrotic syndrome due to membranous nephropahy, focal segmental glomerulosclerosis or minimal-change nephropathy, but classified patients in three subgroups instead of two. Since as many as 59% of membraneous nephropathy patients and 38% of focal segmental glomerulosclerosis patients had a SI lower than the cut-off value of 0.2 but higher than 0.1, we classified proteinuria as highly selective (SI 0.10), moderately selective (SI 0.11 0.20), or nonselective (SI 0.21). Although the total amount of proteinuria was comparable in the three groups, a significant difference in clinical remission, in progression to chronic renal failure and in response to therapy was found Tables 1 and 2. A correlation between the SI and the semiquantitative evaluation of histologic lesions showed that the extent of tubulointerstitial damage in patients with highly or moderately selective proteinuria was significantly less severe than in those with nonselective proteinuria, while global glomerular sclerosis did not show a significant relationship with SI Table 3. Bakoush et al¹³³ have very recently confirmed in a group of 56 proteinuric patients with various glomerular diseases that the IgG/transferrin SI below or above 0.2 is significantly correlated with the severity of histologic lesions Table 3. The same group of investigators²⁶ has proposed a different measure of SI, based on the comparison of the clearance of IgM (instead of IgG) to that of albumin (instead of transferrin), on the assumption that the higher molecular weight of IgM in comparison with IgG could reveal a more severe disruption of the permeability of the glomerular capillary wall. Dividing 67 patients with various primary glomerular diseases (IgA nephropathy, membraneous nephropathy and mesangial proliferative glomerulonephritis) into two subgroups, respectively, with low (0.002) and high (>0.002) SI based upon IgM (IgM-SI), they found that at the end of the follow-up significantly less patients with low IgM-SI had a decline of creatinine clearance, despite the fact that the initial average amount of proteinuria was higher in patients with low IgM-SI Table 2. In patients with low IgM-SI, the number of obsolescent glomeruli was significantly lower and the extent of interstitial fibrosis significantly less marked Table 3. No significant difference could be found when the same parameters were correlated with albuminuria. At the multiple regression analysis, IgM-SI, but not albumin excretion rate, appeared to be significant predictor of decline of renal function Table 2.

Table 1 - Predictive value of some qualitative markers of proteinuria for clinical remission in primary chronic glomerulonephritides.

Full table

Table 2 - Predictive value of some qualitative markers of proteinuria for progression to chronic renal failure in primary chronic glomerulonephritides.

Full table

Table 3 - Correlations between some qualitative markers of proteinuria and severity of histological damage in primary chronic glomerulonephritides.

Full table

Since the discriminant factor, in the evaluation of the SI, is the urinary excretion of proteins with the highest molecular weight, the urinary excretion of IgG has been used as a marker of severity of the glomerular damage by some investigators. Reichert, Koene, and Wetzels¹³⁴ evaluated the predictive value for progression of urinary excretion of IgG, together with total proteins, albumin, and transferrin, in 22 patients with membraneous nephropathy, nephrotic syndrome, and normal renal function. Only excretion of IgG and transferrin, but not that of total proteins, was significantly different in progressors versus nonprogressors. By multivariate analysis, only IgG excretion was independently associated with deterioration of renal function Table 2.

We studied the predictive value of the excretion of IgG, expressed in milligrams per gram of urinary creatinine and logaritmically transformed, in 78 patients with biopsy-proved membraneous nephropathy¹³⁵. The excretion of IgG, but not albumin, transferrin, or 24-hour proteinuria, significantly correlated with the severity of the tubulointerstitial damage, while no correlation was found between glomerulosclerosis and any of the proteinuric parameters, including IgG excretion Table 3. In one half of the patients, who had a nephrotic syndrome and a normal renal function at the time of the evaluation of proteinuria and renal biopsy, and a sufficiently long follow-up, complete or partial remission was significantly different (100% vs. 20%) according to a urinary excretion of IgG lower or higher of a cut-off level of 110 mg/g urinary creatinine Table 1, as was progression to chronic renal failure (0% vs. 35%); on the contrary, an amount of 24-hour proteinuria below or above 5 g did not show any discriminant value for either remission or progression Tables 1 and 2. In the same study¹³⁵, 19 patients treated with immunosuppressive therapy were compared with 19 untreated patients. There was no difference in the rate of remission or progression to chronic renal failure between treated and untreated patients when the baseline values of IgG were less than the cut-off value. When IgG excretion values were greater than the cut-off value, however, a difference between treated and untreated patients for progression to chronic renal failure at the limit of statistical significance (10% vs. 60%; P = 0.05) was found.

We have concluded a similar study in 50 patients with primary focal segmental glomerulosclerosis (FSGS)¹³⁶ and in 30 patients with membranoproliferative glomerulonephritis (unpublished observations), with very similar results Table 1,2,3. In 29 FSGS patients with nephrotic syndrome and baseline normal renal function the fractional excretion of IgG, below or above a defined cutoff, predicted remission in 91% and progression to end-stage renal failure in 71% of patients, respectively. Furthermore, a significant difference (P = 0.000) was observed for therapy responsiveness according to fractional excretion of IgG: 70% of patients with fractional excretion of IgG <0.025 were responsive to steroids alone, 80% of patients with fractional excretion of IgG 0.025 to <0.140 were responsive to steroids and cyclophosphamide in combination, 100% of patients with fractional excretion of IgG>0.140 were unresponsive to both drugs¹³⁶.

In our three cohorts of nephrotic patients (membranous nephropathy, focal segmental glomerulosclerosis, and membranoproliferative glomerulonephritis), we have measured the urinary excretion of another HMW protein, ₂-macroglobulin, of molecular radius 90 Å, the concentration of which is usually extremely low, below the detectability limits, even in many pathologic conditions characterized by increased size permeability of the glomerular barrier. This very large protein was detectable in the urines in the totality (focal segmental glomerulosclerosis and membranoproliferative glomerulonephritis) or great majority (79%, in membranous nephropathy) of patients in whom the IgG fractional excretion was above the cut-off levels selected for each of the three diseases, while it was undetectable in the majority of patients whose fractional excretion of IgG, even though markedly increased, was below the cut-off levels. These findings suggest that the large pores permitting the transglomerular passage of HMW proteins in nephrotic patients discriminate between IgG (55 Å) and ₂-macroglobulin (90 Å), as they discriminate, according to Bakoush et al²⁶, between IgG and IgM (120 Å). They can be explained assuming that discriminatory large pores exist, the density of which declines with their size.

Predictive value of fractional excretion of proteins of very LMW ("tubular" proteins)

In 1987, Nagy et al¹³⁷ studied the proteinuria in 45 patients with IgA nephropathy and found that the presence of LMW proteins in the sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) proteinuric pattern was significantly associated with higher values of serum creatinine and was correlated with tubulointerstitial lesions Table 3. In 1991, Woo et al¹³⁸ characterized proteinuria with SDS-PAGE in 60 patients with the same disease and found that the presence of LMW proteins was significantly associated with a higher incidence of chronic renal failure after 6 years of follow-up Table 2; the proteinuric patterns were also correlated with tubular atrophy and tubulointerstitial lesions Table 3. We¹³⁹ measured LMW proteins by SDS-PAGE in 145 patients with proteinuric primary glomerulonephritis (focal segmental glomerulosclerois, membranous nephropathy, and membranoproliferative glomerulonephritis), further distinguishing the molecular weight according to the presence or absence of LMW proteins of a molecular weight below 20 kD (as low as 10 kD). We found that the presence of such proteins of very LMW was significantly associated with a higher serum creatinine level, more marked proteinuria, and more severe tubulointerstitial damage Table 3. The presence at baseline of LMW <20 kD significantly predicted the progression to chronic renal failure Table 2. In the 20 patients of this cohort who were treated with steroids (alone or in combination with cyclophosphamide), response to therapy was significantly lower in those with presence of LMW proteins <20 kD (30% vs. 80%). Reichert, Koene, and Wetzels¹⁴⁰ and our group^135,136 confirmed the value of LMW proteins as sensitive markers of proximal tubular function, and therefore of tubulointerstitial damage and progressive disease, measuring directly the urinary excretion of single proteins of this class. Reichert, Koene, and Wetzels¹⁴⁰ measured the urinary excretion of ₂-microglobulin in 30 patients with membranous nephropathy, nephrotic syndrome, and normal renal function and found that progression to chronic renal failure was significantly more frequent when its excretion was greater than 500 ng/min Table 2. In 48 patients with membranous nephropathy, we¹³⁵ found that the excretion of ₁-microglobulin was significantly associated with the extent of tubulointerstitial damage, global glomerular sclerosis, and arteriolar hyalinosis Table 3. By contrast, no correlation was found between 24-hour total urinary protein excretion, albumin, or transferrin, and any of the histologic features. By multiple regression analysis, in a larger population of 78 patients with membranous nephropathy, we found a significant correlation between urinary ₁-microglobulin excretion and that of IgG, but not of albumin or total urinary protein excretion¹³⁵. A separate analysis, in the subgroup of patients who had nephrotic syndrome and normal renal function at the time of evaluation of proteinuria, revealed that those with a very high urinary excretion of ₁-microglobulin (higher than a cut-off level of 33.5 mg/g urinary creatinine) progressed more frequently to chronic renal failure Table 2 and had a lower rate of clinical remission Table 1. As reported above, in the same population of patients urinary excretion of IgG was also significantly correlated with remission and progression to chronic renal failure, while an amount of 24-hour proteinuria below or above 5 g did not show any discriminant value either for remission or for progression. By multiple logistic regression analysis ₁-microglobulin, but not IgG or 24-hour proteinuria, was significantly associated with progression to chronic renal failure. Only in patients with ₁-microglobulin excretion above the cut-off level, a significant difference in the rate of progression to chronic renal failure was found between treated (steroids and immunosuppressors) and untreated patients (17% vs. 100%). A similar study in 50 patients with primary focal segmental glomerulosclerosis¹³⁶ and in 30 patients with idiopathic membranoproliferative glomerulonephritis (unpublished data) gave comparable results Tables 1,2,3.

CONCLUSION

Two major mechanisms are responsable for the abnormal urinary excretion of proteins that characterizes all glomerular diseases, with or without nephrotic syndrome. The first is an increased charge and size permeability of the glomerular capillary wall, leading to the transglomerular passage of albumin and of proteins of HMW that usually do not cross the glomerular barrier. The second is the consequent impairment of the mechanism of reabsorption of all proteins and, in particular, of the LMW proteins by the epithelial cells of the proximal tubule, due to increased work and/or toxic injury deriving from the increased load of the abnormally filtered proteins in the tubular lumen.

The quantity and the molecular weight and radius of the proteins that reach the tubular lumen increase progressively with the increasing severity of the disruption of the structural integrity of the glomerular capillary walls, with the consequent impairment of the selectivity of the "barrier" to the transglomerular passage of proteins. In the less severe cases with selective proteinuria, albumin (molecular weight = 69 kD; molecular radius = 36 Å) is the prevalent abnormal protein Figure 2b. With the escalating severity of the glomerular lesions, increasing amounts of proteins of larger molecular weight cross the glomerular capillary wall, and the amount of IgG (molecular weight = 150 kD; molecular radius 55 Å) in the tubular lumen increases progressively Figure 2c. In the most severe cases, even proteins like ₂-macroglobulin (molecular weight = 720 kD; molecular radius = 90 Å) and IgM (molecular weight = 900 kD; molecular radius = 120 Å) reach the tubular lumen Figure 2d. It is, therefore, theoretically correct that the SI of proteinuria based on IgM excretion, or the fractional urinary excretion of IgG, may be a better marker of the severity of the damage of the glomerular capillary wall than the overall amount of proteinuria. The clinical studies reported above, and summarized in the Tables 1,2,3, confirm this assumption.

In all the primary glomerular diseases with prevalent damage of the glomerular capillary walls and proteinuria in the nephrotic range (membranous nephropathy, focal segmental glomerulosclerosis, and membranoproliferative glomerulonephritis), urinary excretion of IgG or IgM, but not 24-hour proteinuria, predicted both clinical remission and progression to chronic renal failure and was correlated with the histologic lesions, especially tubulointerstitial damage.

The increased transit in the tubular lumen of large amounts of proteins of HMW induces a progressively increasing damage of the epitelial cells of the proximal tubule and the activation of the network of cytokines and growth factors that induces interstitial infiltration and fibrosis, primarily responsible for the progressive deterioration of renal function. It is, therefore, theoretically justified that the presence of great fractions of proteins of very LMW, which escaped tubular reabsorption, such as ₁-microglobulin (molecular weight = 31 kD) or ₂-microglobulin (molecular weight = 11.8 kD), in the urine may be a more reliable marker of the severity of the tubulointerstitial damage than the overall amount of proteinuria. The clinical studies reported above, and summarized in the Tables 1,2,3, confirm even this assumption.

Considering the great number of factors that intervene during the course of chronic glomerular diseases and that may affect the evolution of the damage and the natural history, it is surprising to find that the amount of HMW and LMW proteins detected in the urine at the beginning of the observation period is a rather accurate predictive marker of the subsequent outcome. This finding suggests that there is a point of irreversibility of the anatomic damage, both at the level of the glomerular capillary wall and of the tubulointerstitium, of which the amount of HMW and LMW proteins is a very reliable indicator. If such "point of no return" has been already reached at the time of the first qualitative characterization of proteinuria, during the subsequent clinical course a reduction in the fractional clearance of HMW and LMW proteins, marking an improvement of the disruption of the permselectivity of the glomerualr filter, can be considered an improbable event.

It has been shown by us¹³⁵ and by others¹³³ that the fractional excretions of IgG, an HMW protein, and of ₁-microglobulin, a LMW protein, are significantly correlated in proteinuric glomerular diseases, probably because an increased transit of the HMW proteins in the tubular fluid exerts an adverse effect on the tubular cells, resulting in an increased urinary excretion of both HMW and LMW proteins. However, the concomitant measurement of both these urinary markers is clinically justified in our opinion, since IgG is more related to the glomerular damage, and ₁-microglobulin to the tubulointerstitial damage. This damage is consequent but not necessarily strictly correlated with the ongoing glomerular injury. The fact, reported above, that the fractional excretion of IgG is a better predictor of clinical remission than fractional excretion of ₁-microglobulin in patients with membranous nephropathy, while the latter better predicts the progression to chronic renal failure than the former, appear to confirm that the predictive role of these two parameters can be different. However, at the regression analysis in membranous nephropathy and focal segmental glomerulosclerosis, their contemporary evaluation did not show any additional value for predicting the clinical outcome in comparison with each of them considered separately (unpublished data). We still measure the fractional excretion of both these markers as a routine in all proteinuric glomerular diseases, and recommend doing so in the clinical practice, especially when no morphologic data from biopsy are available.

The impact of the baseline fractional excretion of IgG (but not of the 24-hour proteinuria) on the therapeutic response to a prolonged cycle of steroids, given alone or in association with cyclophosphamide in our patients with focal segmental glomerulosclerosis, deserves some commentary. Only when the fractional excretion of this marker of HMW proteins was below a defined cut-off limit, was a total or partial response obtained in the majority of patients. A possible explanation of these results is that therapy is efficacious only when the density of largest pores (of which the fractional excretion of IgG is a good marker) is in the normal range or only moderately increased. As observed by Ikegaya et al⁶¹, steroids and immunosuppressors may correct the size barrier defect, only within a limited range of disruption of the integrity of the glomerular capillary wall.

References

References