Portal de Periódicos da CAPES

Albuminuria, Wherefore Art Thou?

2009; American Society of Nephrology; Volume: 20; Issue: 3 Linguagem: Inglês

10.1681/asn.2009010075

1533-3450

George Jarad, Jeffrey H. Miner,

Biomedical Research and Pathophysiology

No concept in kidney physiology raises as much interest and debate as proteinuria. All agree that the glomerular capillary wall is a highly selective barrier that restricts the passage of plasma proteins—thus its moniker the “glomerular filtration barrier” (GFB). Albumin is the most abundant plasma protein, and significant albuminuria is considered “selective glomerular proteinuria,” in contrast to the low molecular weight proteinuria that is classically linked to tubular abnormalities. Although most attention has been focused on GFB abnormalities as being responsible for albuminuria, Comper et al.1 continue to present evidence for a tubular origin, with the latest appearing in this issue of JASN.2 The view they advocate is in stark disagreement with long-accepted dogma of kidney physiology and pathophysiology, but if these investigators are correct, then there would be a major shift in the way proteinuric kidney diseases are viewed and, most important, treated. Inherent in the hypothesis of a tubular origin for proteinuria is the claim that albumin's glomerular sieving coefficient (GSC) is high, at 0.02 to 0.04. This means that 2 to 4% of the albumin molecules subjected to the GFB cross into the glomerular filtrate. This estimate is approximately 50 times higher than the widely accepted GSC of approximately 0.0006.3 The difference between these values is staggering; if the higher value is correct, then it means that in normal rats (GFR of 2 ml/min), 2 to 4 g/d albumin would be filtered, as opposed to only approximately 66 mg/d with the historically accepted GSC. When scaled to humans (GFR of 120 ml/min), 150 to 300 g/d albumin would be filtered; this level of albumin (essentially all of the albumin in the bloodstream) would obviously have to be reclaimed by a very efficient mechanism in the tubules to explain the lack of nephrotic-range albuminuria and negative nitrogen balance in healthy individuals. Indeed, Comper and colleagues4 hypothesize that such amechanism exists, and that albuminuria is caused primarily by defects in tubular uptake of intact albumin rather than by increased leakiness of the GFB. A corollary of the hypothesis is that albumin is not tubulotoxic, at least under normal conditions. In this issue of JASN, Russo et al.2 use two-photon microscopy in living rats to study the early diabetic kidney's handling of fluorescent Alexa568-conjugated rat albumin. Their data are in agreement with their previous results and support their hypothesis. By comparing the fluorescence signals in Bowman's space with those inside the glomerular capillary, they calculate the GSC of Alexa568-albumin to be 0.034, which is not changed in proteinuric diabetic animals. Filtered fluorescent albumin is rapidly taken up by proximal tubule cells (PTCs) in the normal kidney, but, in proteinuric animals, the retrieval pathway is impaired, resulting first in increased peptiduria and eventually in frank albuminuria. Glycemic control in diabetic animals prevents albuminuria by protecting the retrieval pathway in PTCs. Furthermore, the GSC of a 69-kD fluorescent dextran tracer, calculated to be 0.025, was comparable to that of the fluorescent albumin. The half-life of albumin, however, was longer, and only albumin was able to bind to the brush border of PTCs. The latter suggests that the two molecules are processed differently after filtration, consistent with resorption of intact albumin (but not dextran) by PTCs. The authors’ high GSC hypothesis has met with intense criticism, and interested readers are encouraged to consult relevant publications, commentaries, and responses5–8 for details. It is notable that the data of Comper and colleagues have been criticized for potential artifacts as a result of low signal-to-noise ratios in their two-photon analyses of fluorescent albumin levels in plasma and ultrafiltrate in living rats, yet their calculated GSC for FITC-dextran agrees with that determined by Rippe and colleagues,9 providing a sense of validity to their state-of-the-art two-photon method; however, whether fluorescently labeled albumin is a true surrogate for native albumin may be questionable. Although the authors assume that labeling albumin with Alexa568 does not change its physical characteristics, little is known about the ability of Alexa568-albumin to associate with plasma lipids and fatty acids, which may affect its effective molecular radius, shape, deformability, and interactions with the GFB, and is the elapsed time between bolus injection and detection of fluorescence long enough for such associations to occur? In any event, regardless of whether the calculated GSC of 0.034 is applicable to native albumin, an important message of the article is that the GSCs of both Alexa568-albumin and FITC-dextran remain the same (respectively) in normal and proteinuric diabetic rats, suggesting that tubular dysfunction is in fact responsible for the increased albumin excretion in early diabetic nephropathy. This is by no means a novel concept,10 but the direct visualization of quantitative differences in the extent of association of filtered Alexa568-albumin with the PTC brush border by the two-photon method2 provides a powerful demonstration. Another important criticism of the hypothesis of the tubular origin of albuminuria is the lack of a convincing demonstration of the high-efficiency albumin retrieval mechanism that would be required to prevent nephrotic syndrome. The hypothesis is effectively disproved if PTCs are not able to absorb huge quantities of albumin from the tubular lumen and then transport it intact to the circulation. So where is the smoking gun? One plausible reason that Park and Maack did not find evidence for this mechanism in their classic studies11 has already been presented,8 but why is fluorescent albumin confined primarily to the apical pole of PTCs in the two-photon movies and micrographs, where it continues to accumulate at the later time points,2,4 if it must be returned to the circulation through the basal aspects of these cells? Explanations that have been provided are that the process of transcellular albumin transport is too rapid, the vesicles involved are too small to be visualized, and the mechanism responsible has yet to be fully characterized.6 Another possible explanation for the data, if one accepts the high GSC, is that the large filtered load is resorbed but in the form of difficult-to-detect degradation peptides rather than as intact albumin; this would, however, require the plasma albumin half-life to be on the order of hours rather than weeks. As evident at a 2008 American Society of Nephrology Annual Meeting symposium devoted to this topic, the controversy surrounding the notion of a high GSC coupled with a high tubular albumin resorption capacity has evolved into an emotionally charged debate. It seems that a satisfactory resolution may require even more sophisticated techniques and experimental approaches that will provide the proverbial “slam dunk” to end the controversy with indisputable data. As glomerular biologists, we eagerly await a satisfactory mechanistic explanation for how the increasing number of podocyte gene mutations responsible for albuminuria in patients and in animal models can be reconciled with the concept of a tubular origin for proteinuria. Disclosures None. Our research has been supported by grants from the National Institutes of Health (R01DK078314 and R21DK074613 to J.H.M. and P30DK079333 to G.J.), by an Established Investigator Award from the American Heart Association (J.H.M.), and by an Alaska Kidney Foundation-American Society of Nephrology Research grant (G.J.).