Background
Diabetic kidney disease (DKD) is a major complication of type 2 diabetes mellitus (T2DM) with up to 40% of DKD patients progressing to end stage renal disease (ESRD) [
1]. Despite the improvements in the treatment for DKD, the risk of ESRD remains high. Potential new therapy must be evaluated in clinical trials, which are very costly. A successful trial must accumulate enough end-points to give adequate power for detecting a risk reduction between the placebo and the treated group. Biomarkers that can identify and target high risk patients for clinical studies or more aggressive therapy would be extremely useful. The paradigm of the natural history of DKD continues to evolve. DKD clearly does not follow the pattern of glomerular hyperfiltration progressing from microalbuminuria to increasing degrees of overt albuminuria and declining GFR in many patients [
2]. Macrovascular disease rather than classical diabetic nephropathy has been increasingly recognized in the pathogenesis of GFR decline in many DKD patients [
3] [
4]. Although, albuminuria is an important marker to diagnose and predict the progression of DKD [
2], nearly half of T2DM patients may have decreased glomerular filtration rate (GFR) before the onset of albuminuria. Conventional biomarkers cannot accurately identify T2DM patients at higher risk of rapid GFR decline and new biologic markers that can predict DKD progression are required.
DKD has been viewed traditionally as a disease of accelerated matrix deposition leading to progressive glomerular and tubulointerstitial fibrosis with the final consequence of ESRD [
2]. More recently, an important role of inflammatory cells such as monocytes/macrophages has been recognized in the pathogenesis of DKD progression [
5]. Monocyte chemoattractant protein-1 (MCP-1), a member of the CC chemokine family promotes macrophage accumulation both in animal models and several renal diseases [
6‐
8]. Increased urinary MCP-1 levels have been shown to predict adverse outcomes in proliferative kidney diseases such as lupus nephritis [
9]. Previous investigations have shown that the renal MCP-1 expression is also elevated in DKD. Urinary MCP-1 levels correlated with the degree of tubulointerstitial leucocyte infiltration [
10,
11], but at present, its role in predicting renal prognosis in DKD remains unclear.
The balance between protective growth factors and pro-inflammatory cytokines likely determines the degree of renal tissue damage and disease progression in DKD. In contrast to MCP-1, epidermal growth factor (EGF), a peptide growth factor probably has a protective role during kidney injury. EGF expression within the kidney is decreased in several kidney diseases [
12,
13]. Low urinary EGF levels have been found to be predictive of kidney function decline in non-diabetic renal diseases. Therefore, both urinary MCP-1 and EGF could serve as favorable biomarkers for kidney damage in DKD. Previously, it has been shown that urinary EGF/MCP-1 ratio was a better prognosticator of long term outcome compared to either cytokine alone in IgA nephropathy [
14]. Currently, few studies have evaluated the roles of MCP-1 or EGF in predicting DKD progression across a broad spectrum of kidney function and none have evaluated the role of their ratio. This study aimed to test the hypothesis that baseline levels of urinary MCP-1 and EGF or their ratio would predict rapid decline of estimated GFR in a cohort of T2DM patients with CKD independent of conventional clinical risk factors. We also evaluated the biomarkers in subgroups with high cardiovascular risk and those low albuminuria to explore the potential value of these biomakers in subjects who may have non-classical mechanisms for DKD progression.
Discussion
The ability to identify which DKD patients will develop rapid decline in kidney function is essential and currently available clinical tests cannot predict renal outcomes accurately. The present study indicate that MCP-1, EGF, and the EGF/MCP-1 ratio represent potential biomarkers of rapid progression in DKD. UMCP-1, and UEGF/MCP-1 ratio showed significant predictive power after multivariate analysis with conventional factors (blood pressure, GFR and UACR). This might suggest that these biomarkers would not be simple surrogate indexes of baseline estimated GFR, but potential markers on their own, predicting DKD progression beyond the conventional risk factors.
The recruitment of inflammatory cells into renal tissue has a pivotal role in the progression of various renal diseases by promoting a microenvironment that amplifies tissue injury and fibrosis [
20,
21]. MCP-1-mediated macrophage accumulation and activation are critical events in the development of diabetic renal injury in animal models [
20,
21]. MCP-1 protein and mRNA were detected in cortical tubuli, and infiltrating mononuclear cells in the kidneys of patients with DKD [
10], Urinary MCP-1 levels correlated with the severity of both the tubulointerstitial and glomerular lesions. Other studies have also shown correlations between urinary MCP-1 with baseline proteinuria and renal function in DKD [
10,
22,
23]. In line with previous studies, we also found that UMCP-1 strongly correlated with the level of albuminuria. In addition, we also found that high UMCP-1 was a predictor of rapid GFR loss in DKD. The relationship between urinary MCP-1 levels and subsequent GFR decline had been observed in earlier studies. In a small study, urine MCP-1 levels was found to correlate with the rate of renal function decline in DKD patients over a 6 year period [
24], but the authors did not adjust for conventional risk factors in the study. Recently, urinary MCP-1 was found to be independently associated with the rate of GFR decline in a Canadian cohort with advanced stage 3 to 4 DKD [
25]. Our study extended previous observations by showing that high UMCP-1 was predictive of rapid renal function loss across a broad spectrum of kidney function that was independent of conventional factors in Asian patients with stage 1 to 5 DKD.
EGF plays important role in restoring barrier function in the healing phase of renal injury and is also a critical tubular cell survival factor [
26,
27]. A role for EGF in predicting outcome in chronic kidney diseases is being increasingly recognized, but there is limited data in DKD. Urinary EGF levels have been shown to correlate with the severity of tubulointerstitial fibrosis in primary glomerulonephritis [
13], and low urinary EGF has been shown to be a risk predictor of kidney progression in non-diabetic kidney diseases [
28]. More recently, urine EGF was identified as a marker for DKD by urine peptidomic profiling of a diabetic rodent model [
29], and reduced urinary EGF levels had been reported in patients with DKD [
30,
31]. In this study, we demonstrated the predictive role of low UEGF for detecting rapid kidney decline across a broad spectrum of GFR and albuminuria including a subgroup of patients with low albuminuria (mostly in the microalbuminuria range). Previously, the predictive role of urinary EGF has been investigated in diabetic patients with preserved GFR and normoalbuminuria [
29]. The investigators found that low urinary EGF was associated with rapid decline in renal function and predicted incident CKD independent of standard risk factors. DKD is associated with many alterations in the structure of glomeruli, tubulointerstitial and vascular compartments [
32]. In line with previous studies, we showed that UEGF correlated with GFR, but not with albuminuria consistent with a predominant role of UEGF as a marker of renal tubulointerstitial involvement than of glomerular damage [
13,
33]. However, unlike the previous study of patients with preserved GFR and normoalbuminuria [
29], the utility of urine EGF to predict rapid decline in GFR was not apparent after adjustments for albuminuria in our study. Overall, this data is consistent with the hypothesis that EGF may be important in the pathogenesis of DKD and low UEGF may be a useful marker for progression especially in patients with normo- or low albuminuria. It is possible that the benefit of UEGF as a biomarker of rapid progression over traditional markers may be diminished in the context of more advanced DKD or marked proteinuria.
DKD is typically characterized by a high prevalence of cardiovascular risk factors and increased risk for cardiovascular events [
2]. Macrovascular disease has been proposed as a mechanism for decreased GFR in some DKD patients [
3,
4]. Biomarkers that can identify either cardiovascular or renal outcomes in patients with high cardiovascular risk would be especially useful as interventions may have the greatest benefit in this group [
2]. We explored the role of the biomarkers specifically in a subgroup with high cardiovascular risk defined as > 10% probability for developing cardiovascular event at 10 years based on the Framingham risk score. Although the proportion with high risk was quite high (44.6%), only 3.7% actually had already had CV events. Thus most patients may not have extensive macrovascular disease at baseline. The proportion with high cardiovascular risk or the Framingham risk score at baseline did not differ between those who developed rapid GFR decline compared to those who did not suggesting that high CV risk might not be a major factor for GFR decline in our study population. None of the clinical parameters or biomarkers were predictive of rapid GFR decline in the high cardiovascular risk subgroup, which probably reflected the small number of patients in the subgroup. Larger studies are necessary to allow firm conclusions on the role of these biomarkers in high cardiovascular risk patients.
Albuminuria is commonly used to assess kidney disease progression among patients with DKD, but some studies have shown that this marker may have insufficient ability to predict kidney disease end points on its own [
19]. On this basis, exploring novel biomarkers that can more reliably identify individuals at risk of experiencing adverse renal outcomes is warranted. Inflammatory cytokines and growth factors are logical candidates for biomarkers of progressive DKD as they have been involved in the pathogenesis of renal fibrosis and are easily measurable in the urine. In our study, albuminuria was still the strongest predictor for rapid GFR decline. The benefit of adding additional biomarkers varied depending on the statistical method used. The C-statistics test showed that UACR had a fairly good discrimination for rapid GFR decline. The addition of other biomarkers did not improve the area under the curve. However, the C-statistics test may be insensitive when adding a new predictor to a model and may occasionally produce incorrect estimates [
34]. This phenomenon is particularly perceptible when the baseline model includes strong predictors especially in the context of low numbers of patients such as in our study. On the other hand, by using the multivariate model, we showed that MCP-1 or EGF/MCP-1 were independent predictors of rapid GFR decline. Together with findings from previous studies, our results suggest that urine MCP-1 could be a promising biomarker for rapid GFR decline in DKD across a broad range of GFR, but additional studies are necessary to confirm the added benefit to conventional factors including UACR. Findings of this study has several implications. Urine MCP-1 might be utilized to identify high risk patients for inclusion into clinical trials or for more aggressive intervention. The findings also provide additional support for the role of MCP-1 in the pathogenesis of human DKD progression. This is especially important as specific inhibitors of MCP-1 are now available for clinical studies with the potential to impact proteinuria [
35]. It would be of interest to see if high urinary MCP-1 could identify those who would mostly likely to respond to MCP-1 blockade. To our knowledge, this is the first study to evaluate the role of EGF in addition to MCP-1 as a ratio in DKD. Earlier studies have shown that combinations of EGF with MCP-1 in the form of EGF/MCP-1 ratio was a better predictor of eGFR decline in non-diabetic kidney disease compared to either biomarker alone [
14]. In this study, although high EGF/MCP-1 was an independent of rapid loss of GFR in DKD, the overall performance of the ratio was comparable to UMCP-1 alone. Therefore, our study does not support the benefit of routine EGF measurement in addition to MCP-1 given the additional costs. EGF or EGF/MCP-1 may be useful in the DKD subgroup with normo- or low grade albuminuria and decreased GFR. This subgroup may constitute a larger proportion of DKD than previously recognized [
2,
4] but larger studies are necessary to evaluate the cost-benefit of EGF measurement to predict adverse renal outcomes in this group.
This study has several limitations. This is a single center study involving small numbers of subjects with a short follow-up period of 24 months. Urine biomarkers were measured in a cohort of high risk of DKD patients who might have had a more rapid decline in renal function in a limited time of follow-up. The inclusion of higher risk patients and single-center nature of the study design may have an impact on its external validity to a broader group of T2DM patients. It remains to be seen if urinary MCP-1 and or EGF really predict future loss of kidney function in more or less stable, comparable patients or if these values are “just” independently associated with more rapid loss of kidney function. In keeping with standard clinical practice, we did not perform kidney biopsies routinely in patients with DKD in the absence of clinical features suggestive of other kidney diseases [
3]. The prevalence of non-diabetic kidney diseases in patients with diabetes reported in the literature is highly variable, but about 10% of unselected patients with early DKD and albuminuria may have non-DKD as an underlying kidney disease [
3]. A number of our patients had low albuminuria. However, a lack of albuminuria may not necessarily preclude structural DKD and almost all histopathologic classes of DKD may be present in such patients [
36]. However, given that our inclusion criteria of poorly controlled hypertension or longer duration of diabetes, other diagnoses such as hypertensive nephropathy or other glomerular diseases could not be excluded. Finally, given the limited time, our study did not collect data of major renal outcomes such as ESRD.