| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Editorials |
Research Institute, The Mary Imogene Bassett Hospital, Cooperstown, NY 13326, Fax 607-547-3061, E-mail theodore.peters{at}bassett.org
A report in this issue by Osicka and Comper (1) points out the complexity of the urinary excretion of albumin in mild diabetes mellitus and raises a challenge to clinical chemists for the early detection of renal damage. The authors demonstrate a hitherto unrecognized form of albumin that is not detected by customary immunoassays of urine.
For background, let us look first at the current understanding of the excretion of albumin by healthy kidneys. The two kidneys together treat 
650 mL of plasma every minute. This means that, in the course of a day, 37 000 g of serum albumin pass through the glomeruli, of which 1.3 g leak through the glomerular walls (2). This is 0.004% of the presented albumin, an amazing efficiency for any filtration system. It is equally amazing that only 15 mg, 1%, of the 1.3 g of leaked albumin appears in the urine daily as determined by RIA.
The remainder is not saved as albumin, but is degraded to fragments by the renal proximal tubules. (In perspective, this renal degradation, 1.3 g, is 10% of the daily albumin turnover of 14 g.) The albumin is taken up into endosomes of the proximal tubular cells. These endosomes then merge with lysosomes, where lysosomal enzymes, cathepsins B and D, and particularly the cysteine proteases (3) cleave the 66 500-Da albumin into 500- to 15 000-Da fragments (2)(4).
Where do these fragments go? A study with isolated human proximal tubular cells (HK2 cells) in culture (5) demonstrated that radiolabeled albumin enters the cells from the apical pole (the lumenal side). Approximately 35% of the resulting fragments appear to be released at the apical pole and 65% at the basal pole (into the renal vein). This means that the majority of the amino acids from the 1.3 g of filtered albumin are conserved by the body. An in vivo study in humans confirmed the presence of radiolabeled albumin fragments in the circulation (4). (The authors reported the amount of fragments in the blood as up to 18% of the amount of fragments found in the urine, but fragments would be rapidly cleared from the blood, whereas they accumulate in urine, so this 18% is a minimal measure of the fragments released into the blood.)
A crucial point is that the fragments of albumin are not detectable by immunoassay, either RIA or immunoturbidimetry. They can be seen on Western blotting (6) (the epitopes on the fragments may be weak or low in number, so that they are detectable only on immobilization), but Western blots are not yet a practical clinical laboratory technique. Thus, laboratories now report only the amount of whole, undamaged serum albumin excreted, 15 mg/day in a healthy person.
In their study, Osicka and Comper (1) measured urinary albumin by three methods: RIA, electrophoresis on polyacrylamide gel (native PAGE), and HPLC using a size-separation GF-250 column. In healthy individuals without diabetes, the results of the three methods agreed well (7). In eight patients with minimal diabetic damage, the HPLC values were invariably higher than the RIA values, with mean concentrations of 44.6 mg/L for HPLC vs 18.0 mg/L for RIA. In four of these patients, the RIA value was <25 mg/L, whereas the HPLC value was >41 mg/L. Hence, approximately one half of the findings would have been reported as normal, but some renal damage was apparently present. In a larger study, 33% of 97 diabetic individuals were found to be normal by RIA but to have increased urinary albumin concentrations as measured by HPLC (of 86 controls, 3.5% had increased concentrations) (7).
The difference between the HPLC and RIA results has been termed "immunochemically nonreactive albumin". In earlier reports, it was called "ghost" albumin (8), a catchier but less descriptive title. The contribution of Osicka and Comper (1) was to isolate the 66 500-Da immunochemically nonreactive albumin from diabetic urine by further purification by HPLC after removal of immunochemically reactive albumin on a column of immobilized anti-human albumin antibodies. This nonreactive albumin form showed normal size and migration on HPLC and PAGE, but broke down completely into fragments when tested by reducing sodium dodecyl sulfate (SDS)-PAGE [see Fig. 4 of their report in this issue of the journal (1)]. Hence, immunochemically nonreactive albumin was found to be an intact, although damaged, whole albumin molecule. Extensive "nicking" was apparently the result of processing by renal lysosomal enzymes, but the albumin molecule was still held together by the 17 disulfide cross-links of its chain. On reduction in SDS-PAGE, these bonds were broken and the fragments were released. The heavily nicked albumin molecule is not immunoreactive and so is missed by RIA and immunoturbidimetry.
It has been recognized that an early feature of diabetic nephropathy is impairment of the lysosomal mechanism for degrading albumin and other proteins (4)(9)(10). Increased filtration of albumin at the glomerulus is not a major factor at this stage. One proposal is that an early event is the increase in "large pores" in the glomerular capillary wall, which allows more high-molecular-weight proteins (e.g., IgG and IgM) to escape. These saturate the lysosomal degradative mechanism so that digestion of absorbed proteins is less complete.
In the case of immunochemically nonreactive albumin, cleavage of the albumin chain appears to proceed relatively normally, but release of the fragments is impaired. This might mean that an early effect on the proximal tubular lysosomes is a decrease in their reducing capacity, a loss of RSH groups, which are normally in abundance. Hence, a redox change in the climate of the lysosomes could be the cause of escape of intact but nicked albumin. This is an interesting area for investigation.
For clinical chemists, the challenge is to devise practical methods to measure the total intact albumin excreted, whether undamaged or extensively nicked by proteases, as the first indication of diabetic renal damage. HPLC is not readily usable in the routine clinical laboratory, but perhaps capillary electrophoresis or improved agarose or other thin-layer gel electrophoresis techniques could be applied to the assay of urinary proteins. New reference interval values would be required, but the benefit may be detection of incipient damage several years earlier than with current immunochemical methods. The detection of immunochemically nonreactive albumin was found to precede increases of RIA values by a mean of 3.9 years in type I diabetes and 2.4 years in type II diabetes (11). A great deal of further study is needed, but it should be worth the effort.
Acknowledgments
I would like to thank Ann Eldred, MD, and Valerie Bush, PhD, for helpful suggestions regarding the manuscript.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
