Background
Chronic kidney disease (CKD) is a global public health issue [
1]. The prevalence of CKD is estimated to be 8 to 16% worldwide [
2]. A considerable proportion of CKD patients eventually progresses to dialysis-dependent end stage renal disease (ESRD), which is an important economic burden to the health care system [
1]. Given the increasing prevalence of CKD, the associated burden of CKD is heavy and ever-rising worldwide [
3,
4].
CKD is characterized by progressive renal function loss irrespective to the primary cause of kidney damage [
5]. However, the underlying mechanism of progressive renal function loss in CKD, it still remains incompletely understood. Emerging evidence shows that mitochondrial dysfunction plays an important role in this regard, especially in the setting of diabetic nephropathy [
6‐
8]. Glomerular injury or intra-glomerular hypertension results in proteinuria, which is taken up by renal tubular cells, resulting in the generation of reactive oxygen species and cellular oxidative stress [
9,
10]. The end result is mitochondrial damage, which contributes to the tubular cell damage as well as epithelial-mesenchymal transition (EMT) [
11,
12]. In experimental models of CKD, aborting mitochondrial dysfunction prevents renal tubular cell EMT [
9] and renal fibrosis [
13].
It has long been known that extracellular cell-free mitochondrial DNA (mtDNA) could easily be detected in various body fluids. Previous studies indicated that mtDNA is released into the systemic circulation after mitochondria is damaging [
14,
15]. In addition to blood, dysfunctioned mitochondria in renal tubular cells are also found to release mtDNA into urine, and extracellular mtDNA in urine has recently been explored as the biomarker of various kidney diseases [
16‐
21]. Our previous study showed that urinary mtDNA levels are associated with the severity of renal impairment and histological scarring in diabetic kidney disease [
22]. In the present study, we determine the relation between urinary mtDNA level and kidney damage in non-diabetic CKD.
Discussion
In this study, we show that mtDNA is easy to detect in the urinary supernatant of patients with non-diabetic CKD. Urinary mtDNA level correlates with the degree of renal dysfunction and histological damage. Although urinary mtDNA levels are associated with renal outcome by univariate analysis, the result was not statistically significant after adjusting for clinical and pathological confounders. The role of urinary mtDNA level as a biomarker of patients with non-diabetic CKD may be limited.
Intrarenal oxidative stress plays a critical role in the initiation and progression of kidney disease [
29]. The re-absorbing process in renal tubule is highly energy consuming, so that intra-renal hypoxia is an early manifestation of kidney injury [
30]. The end result is mitochondrial dysfunction and damage, which release mtDNA into the extracellular space and systemic circulation. In this regard, Chien et al. [
31] previously showed that diabetic patients had significantly less cellular mtDNA content, and diabetic atherogenesis is associated with decreased amounts of mitochondrial DNA. Mishra et al. [
32] further showed a strong relation between peripheral blood mtDNA damage and diabetic retinopathy, and circulating mtDNA level is a noninvasive biomarker of diabetic retinopathy.
In recent years, urinary cell-free mtDNA level has been explored as a biomarker of kidney diseases. Several studies in the setting of acute kidney injury (AKI) showed that urinary mtDNA level is increased in mice [
26] and human [
33] after renal ischemic injury, and the level of urinary cell-free mtDNA correlates with the duration of ischemia. Published data on urinary cell-free mtDNA level in the setting of CKD are few. Our previous study showed that reduction of intra-renal mtDNA level in diabetic nephropathy correlates with the degree of renal damage [
22], suggesting that intra-renal mitochondrial dysfunction is an important feature of diabetic nephropathy. In another study on biopsy-proved hypertensive nephrosclerosis and IgA nephropathy, we found that urinary mtDNA level correlates with the rate of renal function decline and predicts the risk of doubling of serum creatinine or need of dialysis [
34]. In the present study, we extend our observation to non-diabetic kidney diseases with no specific pathological diagnosis.
Our present study showed a significant correlation between urinary mtDNA level and serum creatinine as well as proteinuria. On the first glance, the result may seem counter-intuitive because higher eGFR is expected to be associated with lower proteinuria. However, a lower eGFR probably indicates more permanent renal parenchymal damage and therefore a lower capability to leak mtDNA into the urine. Among the patients with renal biopsy, we observed significant inverse correlations between urinary mtDNA level and the severity of pathological scarring, both in terms of glomerulosclerosis and tubulointerstitial fibrosis. The result is expected because, with increasing fibrosis and declining number of renal parenchymal cells, it is logical to predict that the amount of mtDNA that could be leaked into the urine would reduce. In other words, urinary mtDNA level could be considered as a surrogate marker of the degree of permanent renal parenchymal damage or tissue fibrosis, which also explains its relation with renal survival. Since histological assessment of renal fibrosis is both invasive and subjected to sampling error, our result suggests that urinary mtDNA level may serve as a valuable biomarker of renal damage.
The result of this study is somewhat different from our previous ones [
22,
34]. In our previous study on diabetic nephropathy [
22], urinary mtDNA levels inversely correlate with eGFR and positively with tubulointerstitial fibrosis. In another study, urinary mtDNA levels positively correlate with proteinuria and inversely with decline of eGFR as well as renal survival in patients with hypertensive nephrosclerosis and IgA nephropathy [
34]. The pattern of correlation of our present study is different from the previous one on diabetic nephropathy [
22]. Although the underlying reason is not clear, it is possible that high blood glucose causes metabolic stress and directly triggers the damage to tubular epithelial cells, resulting in the excretion of mitochondrial fragments to the extracellular space and urine. Multiple studies showed that hyperglycemia affects renal tubular cells by several mechanisms: overproduction of reactive oxygen species (ROS), activation of apoptotic pathway, and initiation of autophagy [
35‐
39]. Mitochondria, the powerhouse of all cells, are the center of these events. Urinary mtDNA level is conceivably a surrogate marker of mitochondrial repair and regeneration.
There are other limitations of our study. First, the proportion of patients with biopsy is small, limiting the statistical power of multivariate analysis. There may also be unrecognized selection bias for patients to undergo kidney biopsy. Further studies with a larger sample size from an unselected cohort would be necessary to validate our result. Furthermore, the variability of urine mtDNA level is substantial, which may limit the clinical application of urinary mtDNA for risk stratification of CKD.
In conclusion, our study shows that mtDNA is readily detectable in the urinary supernatant of non-diabetic CKD patients. Urinary mtDNA levels have no significant association with the rate of renal function decline in non-diabetic CKD, although the levels correlate with baseline renal function, proteinuria, and the severity of histological damage. Urinary mtDNA level may be a surrogate marker of permanent renal damage in non-diabetic CKD.