Blood levels of two receptors that bind to tumor necrosis factor–alpha (TNFα) strongly predict early and late stages of kidney function decline in patients with type 1 and type 2 diabetes, reported two studies published online ahead of print by the Journal of the American Society of Nephrology.
Type 1 diabetes patients with the highest concentrations of tumor necrosis factor receptor 2 (TNFR2) had a 60% 12-year cumulative incidence of chronic kidney disease (CKD) Stage 3 or higher, while the 12-year cumulative incidence of end-stage renal disease (ESRD) was 54% for type 2 diabetes patients with the highest concentrations of tumor necrosis factor receptor 1 (TNFR1).
“Before, nobody considered that these markers are useful for diagnostic purposes, and now it looks like they are the best,” said lead investigator Andrzej S. Krolewski, MD, PhD, Head of the Section on Genetics and Epidemiology at the Joslin Diabetes Center and Associate Professor of Medicine at Harvard Medical School.
“You can just measure their levels and predict what will happen within six to 12 years in the future. None of the other markers is able to do this.”
While previous studies showed that the risk of developing diabetes is linked to levels of proinflammatory cytokines, such as TNFα, it was not known whether chronic inflammation also contributes to diabetes-related kidney decline.
“We have been on a long quest for better biomarkers that predict clinically meaningful outcomes in kidney disease,” said Katherine R. Tuttle, MD, Executive Director for Research at the Providence Sacred Heart Medical Center & Children's Hospital in Spokane, WA, and Clinical Professor of Medicine at the University of Washington School of Medicine in Seattle, who was not involved in the studies.
“This is a major breakthrough in the field, and it's the strongest data that we've seen on a new biomarker in quite some time.”
Type 2 Diabetes
Dr. Krolewski has had a long-standing interest in kidney complications associated with diabetes. Based on the past research linking proinflammatory cytokines with type 2 diabetes, he decided to use stored blood samples to examine whether markers of inflammation could predict renal decline.
One of the studies included 410 patients with type 2 diabetes who had been recruited to the Joslin Study of the Genetics of Type 2 Diabetes and Kidney Complications between 1991 and 1995. At the end of 2004 (after eight to 12 years of follow-up), 267 of the patients (65.1%) remained alive without ESRD; 59 patients (14.4%) developed ESRD, 51 of whom died; and 84 patients (20.5%) died without ESRD.
The baseline concentrations of TNFR1, TNFR2, and TNFα were significantly higher in ESRD patients than in patients who remained alive without ESRD, but the effects of TNFR1 and TNFR2 were much stronger than that of TNFα after controlling for clinical covariates.
Moreover, the concentrations of TNFR1 and TNFR2 predicted cardiovascular disease and mortality from all causes, but these effects were smaller than those seen with ESRD. By contrast, ESRD was not associated with high blood levels of two markers of systemic inflammation—interleukin-6 and C-reactive protein.
Most ESRD cases occurred in the quarter of patients with the highest concentrations of TNF pathway markers, with few cases among the half of patients with the lowest concentrations. The 12-year cumulative risk of ESRD was 54% for the quarter of patients with the highest TNFR1 levels but only 3% for the other patients.
“The strength of association was remarkable, especially with controlling for other risk factors,” Dr. Tuttle said in a phone interview.
Among the quarter of patients with the highest concentrations of TNFR1, ESRD was more likely to develop several years sooner in patients with proteinuria than in patients with normoalbuminuria or microalbuminuria.
Regardless of patients' albumin levels, the concentration of TNFR1 was a major determinant of the time to ESRD onset, and it was an even better predictor of ESRD than was albumin excretion rate.
Type 1 Diabetes
Because TNFR1 and TNFR2 levels predicted the development of ESRD in patients without proteinuria, Dr. Krolewski and his team examined whether these receptors contribute to early stages of renal decline in patients with type 1 diabetes who participated in two prospective Joslin Kidney Studies.
Subjects in the first study were recruited from 1991 to 1992 and were followed for eight to 12 years, and participants in the second study were recruited from 2003 to 2005 and were followed for five to seven years. The combined 628 patients had high normoalbuminuria or microalbuminuria and normal renal function, as indicated by glomerular filtration rates.
Over a 12-year period, 69 patients developed Stage 3 CKD or higher. The 12-year cumulative risk of Stage 3 CKD or higher was 60% for the quarter of patients with the highest baseline concentrations of TNFR2 and 5% to 19% for the other patients.
The effect of TNFα did not remain significant after adjusting for clinical covariates. While many baseline clinical covariates were associated with risk for CKD Stage 3 or higher in univariate analysis, all of them except for hemoglobin A1c, albumin excretion rate, and estimated glomerular filtration rate lost significance in multivariate analysis.
Although TNFR1 levels most strongly predicted ESRD and TNFR2 levels most strongly predicted CKD, the concentrations of the two receptors were highly correlated, and their effects were similar.
“At this moment, we don't have the foundation to say that one is more important than the other,” Dr. Krolewski said.
The studies were supported by the JDRF, the National Institutes of Health, the American Diabetes Association, the Uehara Memorial Foundation, and CONACYT Fundacion Mexico en Harvard.
One major strength of the studies was the careful, long-term follow-up of patients, said Frank C. Brosius III, MD, Chief of the Division of Nephrology and Professor of Internal Medicine and of Molecular & Integrative Physiology at the University of Michigan, who was not involved in the studies.
The findings should be validated in a multicenter, prospective study of other diabetic patient populations, though, he added. For instance, all the study participants with type 2 diabetes resided in Massachusetts, and 85% of them were Caucasian.
“We don't know if the findings really apply to everybody around the world,” Dr. Brosius said in a phone interview.
Dr. Tuttle agreed that the findings should be tested in other groups of patients with diabetes and noted that the underlying mechanisms linking TNF receptor levels to kidney decline should be examined.
“Before we start thinking that we should apply them as clinical markers, those sorts of studies need to be done,” she said.
The results point to a specific relationship between TNF receptor signaling and renal decline in patients with diabetes, rather than to a process mediated by general inflammation, the authors concluded. They also suggested that systemic TNF receptors, not those located in the kidneys, contribute to the increased risk of early renal decline because concentrations of the receptors in the urine did not correlate strongly with concentrations in the bloodstream.
“That remains to be worked out on an experimental level,” Dr. Tuttle said.
However, an understanding of the underlying mechanisms may be hindered by the lack of a good animal model for studying diabetes-related kidney complications, Dr. Krolewski said. Nonetheless, he is currently conducting experiments to learn more about what is causing elevated levels of TNF receptors and whether they directly damage the kidney or indicate the presence of other factors that cause kidney damage.
“We may want to know right away the mechanism for the association between TNF receptors and end-stage renal disease, but it may take some time to figure out,” he said in a phone interview.
Another potential barrier to clinical implementation is the reliance on the cumbersome enzyme-linked immunosorbent assay (ELISA) for measuring TNF receptor concentrations. Dr. Krolewski is planning to work with biotechnology companies over the next few years to improve the convenience of this test by incorporating it into existing diagnostic platforms.
If TNF receptors are adopted as biomarkers in a clinical setting, they could reveal which patients should receive closer attention to ensure the appropriate management of conventional risk factors for kidney disease, Dr. Tuttle said. For instance, high-risk patients could control their blood glucose levels and take angiotensin receptor blockers or angiotensin-converting enzyme inhibitors to lower their blood pressure.
But these drugs, which are meant to protect renal function, have limited effectiveness, Dr. Krolewski said, adding that TNF receptors can be used as biomarkers to identify suitable patients for clinical trials testing novel kidney protectants.
A test that measures TNF receptor levels “will have the greatest potential if we have some effective intervention,” he said.