Peptide and non-peptide biomarkers of mortality in patients with advanced chronic kidney disease

Abstract: Over the last decade, multiple studies have attempted to identify better risk prediction tools specific to the chronic kidney disease (CKD) population. In many cases, these studies have evaluated the use of various peptide biomarkers as prediction tools. This is common practice in clinical medicine, but in CKD patients these studies are complicated by the fact that most small-to-medium sized compounds are thought to be at least partly metabolized in the kidneys. In fact, a majority of polypeptides smaller than 80 kDa circulate in elevated concentrations in these patients and appear to rise contemporaneously with a drop in glomerular filtration rate (GFR). Thus, elevated levels of many biomarkers may reflect not only increased production as in other patient populations, but also a prolonged half-life due to impaired clearance in the proximal tubules. In Study 1, we analyzed the mortality risk associated with an elevated level of plasma pentosidine in 746 patients with CKD stages 1 - 5 including patients undergoing hemodialysis and peritoneal dialysis treatments. Our main finding is that pentosidine is markedly elevated in CKD and associates with low GFR, oxidative stress and inflammation, but it remains a statistical predictor of mortality in CKD patients also after correction for these confounders. In Study 2, we investigated biomarkers in 543 CKD stage 5 patients initiating dialysis treatment in terms of their ability to classify patients with clinically overt cardiovascular disease (CVD) at baseline and predict all-cause mortality during follow-up for median 28 months. Following adjustments for confounders such as renal function, none of investigated biomarkers were better than simple demographics in predicting adverse outcomes. In Study 3, we again assessed the usefulness of biomarkers, now in 224 prevalent hemodialysis patients. We assessed their ability to classify presence of clinically overt CVD at baseline, as well as predict CVD-related and all-cause mortality during a median 41 months follow-up. We found that while circulating pro-brain natriuretic peptides and insulin-like growth factor 1 predicted baseline CVD, interleukin 6 predicted CVD-related mortality during follow-up, while all-cause mortality was predicted by levels of 8-hydroxy-2’-deoxyguanosine. Thus, no biomarker was useful for all three outcomes, and none of the 13 was statistically better than demographic data in determining baseline CVD or mortality risks in these patients. In Study 4, renal biopsies from 15 patients with glomerulonephritis and three biopsies from healthy living donors were analyzed for expression of the proximal tubular endocytotic proteins megalin and cubilin. Using both immuno-electron microscopy and standard immunohistochemistry, we show that megalin and cubilin are not just two components of a receptor-endocytosis complex, but rather appear to play distinct roles during albuminuria, as evidenced by their differing expression levels. To gain further understanding, we generated a proteinuric zebrafish model and showed that megalin and cubilin indeed are influenced by tubular albumin levels.