Depending on the definition of chronic kidney disease (CKD) employed, there are as many as 20 million Americans currently afflicted with renal disease who are at risk of end-stage renal disease (ESRD). The Dialysis Outcomes Quality Initiative (DOQI), developed by a National Kidney Foundation (NKF) Clinical Practice Guidelines consensus panel in 2002, states that CKD is present when the glomerular filtration rate (GFR), usually measured by creatinine clearance, declines to <90 ml/min/1.73m(2) for at least three months. CKD, in this context, can be subdivided into five stages based on GFR, a broadly accepted renal function test(1) as listed in Table 1.

Table 1. Stages of Chronic Kidney Disease (CKD).

Stage	Description	Glomerular Filtration Rate GFR (mL/min/1.73 m2
1	Kidney damage with normal or >GFR	GFR 90
2	Kidney damage with mild <GFR	GFR 60-89
3	Moderate <GFR	GFR 30-59
4	Severe <GFR	GFR 15-29
5	Kidney failure	GFR <15 or dialysis

Estimating GFR

The determination of GFR by creatinine clearance (CCr) requires a timed urine collection (overnight or 24-hour) and a blood sample, a complicated technique for large scale studies. Instead, CCr is effectively estimated from a single measurement of serum creatinine using the Cockcroft Gault formula.(2)

The Third National Health and Nutrition Examination Survey (NHANES III) used serum creatinine levels to estimate the prevalence and distribution of CKD in the United States (by calculated, estimated GFR and persistent albuminuria). Of 15,625 noninstitutionalized adults aged 20 years and older, the prevalence of CKD in the US adult population was 11% (19.2 million). Distribution of CKD by NHANES III stages is listed in Table 2.

Risk Factors for CKD

Aside from hypertension and diabetes, age is the key predictor of CKD, and 11% of individuals older than 65 years without hypertension or diabetes had stage 3 or worse CKD. One difficulty with NKF criteria for CKD is that the entrance level GFR of 90 ml/min is set too high: all live kidney donors would, by this definition, be immediately placed in the category of having CKD following their voluntary 50 percent loss of GFR. A better, more realistic, definition of CKD would start at a GFR of 50 ml/min.

Furthermore, the DOQI criteria are inappropriate for the elderly. Because the normal aging process is characterized by a 10 ml/min reduction per decade, by the age of 60 almost every healthy elderly individual would, according to the DOQI standards, need to be classified as having CKD.(4)

The cost of ESRD care, as recently stated in the 2003 Annual Data Report of the United States Renal Data System (USRDS),(5) amounts to $22.8 billion, a 33 percent increase over the past eleven years, consuming 6.4 percent of the entire Medicare budget of $242 billion. Advancing age is an epidemiologic correlate of ESRD incidence, while diabetes is the leading cause. Given that the US population is both aging and suffering a pandemic of diabetes, there is an obvious clarion call to "do something" to slow the expanding rolls in dialysis units and on waiting lists for cadaver donor kidneys.

Listed in Table 3 are correlates, identified factors and suspected factors that have been linked with ESRD and might be the focus of interventional strategies to block its progression. Clearly, age, genetic history and ethnicity are beyond change, until the era of gene manipulation medicine is upon us.

In our continuing Cyberounds^® Nephrology series on chronic kidney disease, we examine several purported independent causes and accelerants of CKD as possible targets for action in contemporary care. In this Cyberounds^® we focus on proteinuria, a laboratory sign that, like hematuria, has been recognized as "ominous" from the earliest reports of kidney disease.

Protein Trafficking

Current, ongoing and intense debate questions whether proteinuria is a consequence or cause or both ("chicken and egg") of progressive kidney disease. There is no doubt that proteinuria serves as a reliable marker of renal disease progression.(6) Since the early 1970s, "fixed" (persistent) and increasing proteinuria, are reliable adverse prognostic signs in both diabetic and non-diabetic nephropathies. Large clinical trials(7),(8) found that decreasing the daily quantity of proteinuria with the use of angiotensin converting enzyme inhibitors (ACEIs) and angiotensin-1 receptor blockers (ARBs) slows renal injury and improves patient outcome.

The degree of reduction of proteinuria is important. Without a decrease in proteinuria ACEIs and ARBs have no beneficial effect on progression of renal disease. This 'renoprotective' effect of ACEIs and/or ARBs(9),(10) is distinct from blood pressure reduction whether in diabetic or nondiabetic proteinuric nephropathies. Whether proteinuria per se is only a marker of injury, or also contributes to the kidney damage, is not conclusively established. Partisans in the debate, including a renowned professor of medicine, advocate increasing doses of ACEIs to as large as 1,000 mg/day of enalapril. Because of the importance of this controversy to clinical nephrologists, we now review the evidence indicting proteinuria as an independent risk factor for progressive renal injury.

Proteinuria -- An Independent Risk Factor?

Abbate et al.(11) subjected rats to 5/6 nephrectomy and followed the course of three groups of four rats:

1. normal rats;
2. rats, with proteinuria attributed to altered glomerular permselectivity, increased protein filtration and compensatory increased function of proximal tubular cell; and
3. rats with passive Heymann nephritis, a model of immune-mediated membranous nephropathy.

Cohorts were compared by time course and pattern of protein accumulation and interstitial inflammation.

Reabsorption droplets were noted in remnant glomeruli in regions that subsequently became sclerotic. These reabsorption droplets contained proteins (albumin and IgG) localized in proximal tubular cells (PKC). Interstitial inflammation with MHC-II positive cells (major histocompatability complex class II; found on macrophages and interstitial dendridic cells) developed at sites of protein deposition. Interstitial inflammation continued long after the inciting injury. Osteopontin, a chemokine, whose secretion attracts mononuclear cells, was also detected at the sites of protein accumulation and cellular infiltration. Osteopontin is upregulated in tubular epithelium in association with mononuclear cell accumulation. In both models of progressive kidney disease, osteopontin was detected throughout tubular cells, at sites of MHC-II positive infiltrates.

These observations led Abbate et al. to postulate that the proximal tubule is the initial site of protein accumulation (seen as IgG or albumin) which promotes migration of inflammatory cells, consisting largely of mononuclear cells. As a derivative consequence, at the same location, renal scarring (sclerosis) develops. The authors theorize that inflammatory cells release cytokines, such as TGF-b, a profibrogenic cytokine, which promote excess matrix accumulation. These findings were interpreted as suggesting that it was the intracellular presence of protein that activated a proximal tubule-dependent pathway of chronic interstitial inflammation. In other words, protein, in itself was a renal cellular toxin.

PKC Receptors

Pertinent to the findings of Abbate et al., Brunskill reviewed mechanisms of albumin uptake by proximal tubular cells.(12) Albumin is thought to be freely filtered across the glomerulus, and then by receptor mediated endocytosis enters tubular cells. Megalin and cubulin are two large receptors on PKC cells that enable this processing. Megalin is a large receptor belonging to the low-density lipoprotein receptor family, whose function is internalization of substances prior to their degradation within lysosomes.

Endocytosis is dependent on cyclic adenosine monophosphate (cAMP) and extracellular calcium, and is regulated by several mechanisms including: protein kinase A, protein kinase C and phosphotidylinositol-3 kinase (PI-3-kinase). Albumin can stimulate PI-3-kinase activity. This potential activity is important because megalin is capable of propagating an intracellular signal cascade that may initiate interstitial inflammation. As noted vide infra, protein kinase C (PKC) incubated with human renal tissue plus albumin is stimulated to undergo increased cellular proliferation.

Burton et al.(13) studied stored kidney biopsy specimens (mean of 20 months) primarily from Caucasian patients -- 10 with membranous nephropathy and 10 from individuals with either minimal change disease or without proteinuria who had been biopsied for microscopic hematuria. Proximal tubular cells were cultured and treated with normal human serum protein; fractionated serum protein divided by molecular weight into four groups, A-D.

Cellular proliferation measured by an assay for histone mRNA, which is specific to the S-phase of the cell cycle, was increased in both patient groups, but not in controls. Incorporation of tritiated thymidine, another marker of cellular proliferation, was increased in cells exposed to unfractionated serum and those treated with fraction C, which contained proteins with molecular weight 40-100 kD, composed mainly of albumin and transferrin. When treated with those proteins separately, there was no increase in cellular incorporation of thymidine. LDH release, representing toxicity to cellular membranes, was increased in the same groups.

An interesting footnote is that negligible levels of platelet derived growth factor (PDGF) were found in fraction C, and high levels in fraction D, which ruled out cytokine induced cell proliferation. In this model, therefore, albumin, the principal constituent of proteinuria, was, in itself, nephrotoxic. Burton et al. inferred that in human proteinuric renal disease there is increased turnover of proximal tubular cells.

Protein and Inflammation

There is evidence that protein overload stimulates release of inflammatory and or vasoactive compounds. Zoja et al.(14) studied protein overload in pig tubular cells in vitro. The marker utilized studied was RANTES (regulated upon activation normal T cell expressed and secreted). RANTES is a nuclear factor kappa B (NF-kB) dependent cytokine, which is chemotactic towards macrophages and T lymphocytes. NFkB is a transcription factor that is present in its' inactivated form in most cells.

When the inhibiting subunit IkB is degraded, the gene is activated and causes up-regulation of several proteins. These activated proteins include monocyte chemoattractant protein-1 (MCP-1), which has strong chemotactic activity for macrophages and T lymphocytes. Zoja et al. questioned whether, in renal tubular cells, RANTES is also upregulated after incubation with bovine serum albumin, or immunoglobulin G (IgG). Indeed, Zoja detected a dose and time dependent increase in RANTES levels as detected by ELISA in both milieus, derived from cells obtained mostly from the nephron's basolateral side, not the apical side. A dose dependent increase in NF-kB activation was demonstrated as early as 30 minutes after tubular cells were exposed to albumin. Two specific NF-kB inhibitors added to the culture significantly decreased RANTES indicating that RANTES is an NF-kB dependent protein.

The major components of inflammation resulting from protein overabsorption by tubular cells are monocytes and T lymphocytes, and this inflammatory response occurs on the side of tubular cells in contact with the interstitium, i.e., the basolateral side. As concluded by Zoja et al., RANTES is secreted mostly by the basolateral compartment of the tubular cells. If RANTES is secreted by tubular cells in response to protein overload in vivo, it may, in part, be responsible for the inflammatory response that induces fibrosis. As a footnote, relevant to subsequent discussion, albumin free of fatty acids was equivalent in effect to normal serum albumin. The inference drawn from this finding is that it was the albumin per se, and not the presence of fatty acids plus albumin in the fractional collection, that was nephrotoxic.

The effect of NF-kB inhibition in vivo was studied in rat renal cortical tubular cells.(15) Rats (n=20) made proteinuric with adriamycin had hypoalbuminemia, heavy proteinuria and lower creatinine clearance compared with saline injected controls -- a model of non-immune, non-inflammatory proteinuric glomerulopathy.

Adriamycin treated rats demonstrated progressive increase of interstitial inflammation and edema that by day 28 resulted in thickening of tubular basement membranes, neoangiogenesis and focal matrix deposition. Decreased tubule cell height, from loss of brush border, and cellular atrophy were also noted. However, renal tubular cell cortical interstitial volume in adriamycin injected rats increased as compared with controls.

The inflammatory infiltrate responsible for increased cell volume consisted of monocytes, detected by ED-1 antigen staining, and was particularly increased at areas of tubular cell atrophy. Glomeruli showed increase in mesangial matrix. NF-kB activation correlated with 24h urine protein excretion and with ED-1 cell infiltration. Treatment with pyrrolidine dithiocarbamate, an antioxidant and NF-kB inhibitor, had no effect on proteinuria, but reduced tubule cell atrophy, interstitial volume and ED-1 infiltration. Activation of NF-kB was again shown to have a contributing role in the pathogenesis of monocyte/macrophage infiltration in non-immune proteinuric injury.

Human proximal tubular epithelial cells from five kidneys excised for either carcinoma or deemed inappropriate for donation were cultured in polarized fashion (re-absorption of filtered proteins occurs in apical cells while cytokines and inflammation occurs in basolateral cells).(16) Burton et al.previously demonstrated PDGF secretion occurs into both apical and basolateral tubular cells under normal conditions. In this study, exposure to serum proteins on the apical side induced PDGF (but not MCP-1) secretion towards the basolateral side, and increased PDGF and MCP-1 secretion by the basolateral side. Tumor necrosis factor alpha (TNF-a), an inflammatory cytokine, was not detected.

To narrow down which serum component was responsible for the increased secretion of these two cytokines, serum was fractionated by molecular weight into four groups, A to D. Fraction C (100-40kDa) containing albumin and transferrin correlated with both PDGF and MCP-1 secretion. When these proteins were added individually to the culture, no difference in cytokine secretion was detected, yet unfractionated serum induced increased cytokine secretion. The authors caution that it may be imprudent to extrapolate the significance of increased levels of PDGF and MCP-1 in an in vitro system to in vivo conditions. Nevertheless, in rat tubular cells MCP-1 mRNA is upregulated when exposed to albumin or transferrin.(17) Since Zoja et al.(14) have shown that exposing pig cells to albumin increases secretion of RANTES, a working hypothesis holds that some combination of proteins or a molecule born by proteins is responsible for cytokine release, and that these cytokines attract tissue injuring inflammatory cells.

The Monocyte Role

Monocytes are increasingly incriminated in the pathogenesis of renal fibrosis, as reviewed by Eddy.(18) Cytokines known to be released by monocytes include: profibrogenic factors such as TGF-b, endothelin-1 and tumor necrosis factor alpha (TNF-a). Inhibitors secreted by macrophages included proteases that degrade extracellular matrix, such as tissue inhibitor of metalloproteinase-1 (TIMP-1) and plasminogen activator inhibitor (PAI). Since monocytes are the main component of chronic interstitial inflammation, and this inflammation is induced by proteinuria, it is reasonable to conclude that proteinuria per se is responsible for proximal tubular cell damage. The direct injury is done by inflammatory mediators.

Damage to proximal tubular cells by filtered proteins is documented in many animal studies in vivo, and in human tubular cells ex vivo. Proximal tubular cells proliferate in response to filtered proteins. In addition to proliferating, these cells demonstrate increased synthesis of several chemokines that recruit monocytes and T-lymphocytes. There is probably activation of transcription factor NFkB that in turn activates proinflammatory genes, such as RANTES, IL-1, IL-2, IL-6, MCP-1, and TNF. Tubular cells, as well as macrophages and fibroblasts, are a source of transforming growth factor beta, (TGF-b), a profibrogenic cytokine. TGF-b induces synthesis of extracellular matrix and inhibits its degradation by proteases such as TIMP-1 (tissue inhibitor of metalloproteinase-1) and PAI-1 (plasminogen activator inhibitor-1). The end result is relentless fibrosis.

Summary

Substantive evidence supports a correlation between proteinuria, declining GFR and deteriorating patient prognosis. Once proteinuria is fixed (more than one month), its reduction by treatment with ACEIs and/or ARBs improves prognosis for preservation of renal function as well as patient survival. While nephrologists formerly viewed the amount of proteins in the urine solely as an indicator of the severity of glomerular and tubular injury, we now believe that proteins filtered through the glomerular capillary, together with other risk factors such as hypertension, produce intrinsic renal toxicity. Secondary processes of reabsorption of filtered proteins by activating intracellular events, including up-regulation of vasoactive and inflammatory genes, induce renal interstitial injury and ultimate irreversible fibrosis. Trafficking in proteins in itself damages the kidney.

Treatment regimens that immediately slow or reverse renal injury (steroids for lupus nephritis, for example) hold dual benefits -- at the same time that they decrease proteinuria, they also constrain continuing renal damage consequent to proteinuria. The future of nephrology may well involve vastly improved renoprotection in which gene replacement is feasible for Alport's syndrome, ACEIs and ARBs will be employed for progressive renal disorders and interstitial fibrogenic reactions will be inhibited in glomerular permselective barrier dysfunction such as inflammatory nephritis.(19),(20)