*2.3. Relation between Kidney Function Parameters and Expression of Selected miRNAs*

While *miR-126*, *miR-146a*, and *miR-150* were expressed in all 100 samples of serum, *miR-29c*, *miR-155*, and *miR-223* were not expressed in 12, 8, and 8 serum samples, respectively. Using the Pearson correlation coefficient, none of the six analyzed miRNAs significantly correlated with the parameters of kidney graft function (Figure 3).

**Figure 3.** Correlations between different microRNAs (miRNAs) and parameters of kidney graft function. Correlation between different miRNAs and (**A**) serum urea concentration (s-Urea), (**B**) serum creatinine concentration (s-Creatinine), (**C**) serum cystatin C concentration (s-CysC), (**D**) estimated glomerular filtration rate with Chronic Kidney Disease Epidemiology Collaboration study formula with s-Creatinine (eGFR CKD EPI creatinine), (**E**) estimated glomerular filtration rate with Chronic Kidney Disease Epidemiology Collaboration study formula with s-CysC (eGFR CKD EPI CysC), (**F**) estimated glomerular filtration rate with Chronic Kidney Disease Epidemiology Collaboration study formula with s-Creatinine and s-CysC (eGFR CKD EPI creatinine CysC), (**G**) measured glomerular filtration rate with Chromium-51-ethylenediaminetetraacetic acid clearance (mGFR 51CrEDTA).

#### **3. Discussion**

Although extensive scientific effort has been focused on developing biomarkers to detect kidney allograft disease processes such as rejection or IF/TA, few assays have moved from the research arena to clinical routine. The obstacle to the successful initiation of their clinical use is the still insufficiently validated specificity for renal pathology. Similar to other molecules and proteins, these biomarkers are also subject to glomerular filtration, tubular secretion, and reabsorption. Many of them, depending on the site of their production and paths of elimination, probably reflect merely kidney function and not disease. The identification of sensitive biomarkers able to reflect the subclinical steps of a pathologic process, such as rejection, is therefore of the utmost relevance to recipients of kidney transplants [23–25].

Our research confirmed previous reports showing significantly higher s-Uromodulin in healthy controls (patients without kidney disease) compared to KTRs, who in general belong to the group of CKD patients [4,5,26]. In line with the findings of other similar research in patients with kidney transplants [6], s-Uromodulin levels in our KTRs decreased stepwise from those with almost preserved graft function to the lowest values in KTRs with pre-dialysis CKD-T. We also showed in this study that s-Uromodulin correlates with all parameters of kidney graft function, the strongest association being observed for eGFR CKD EPI creatinine CysC, followed by the serum level of urea, eGFR CKD EPI creatinine, serum level of CysC, eGFR CKD EPI CysC, serum level of creatinine, and 51CrEDTA. Decreased s-Uromodulin was observed in the earliest stages of CKD-T when other markers, such as serum creatinine and even the precise method of GFR determination with 51CrEDTA, had still not crossed the reference range. This indicates that reabsorption of uromodulin, which is exclusively a product of tubules, is probably compromised from the earliest stages of tubulo-interstitial injury. Reduced s-Uromodulin may therefore be a more sensitive indicator of early kidney graft dysfunction not detected by serum glomerular filtration markers. In view of the simple routine of s-Uromodulin measurement and its low costs, monitoring the dynamics of s-Uromodulin concentration may serve as an accurate, noninvasive predictive biomarker of kidney graft injury and outcome.

Given that the concentration of uromodulin strongly reflects renal function, one would assume that it would never be able to act as a reliable parameter of specific kidney graft pathology. However, exceptions have already been found in the field of specificity for certain pathologies of native kidneys, such as gout [27] or Balkan nephropathy [2]. Accordingly, in the case of kidney transplantation, s-Uromodulin could be especially useful to clinicians aware of the advantages of biomarkers reflecting subclinical tubular injury. Regular s-Uromodulin checkups in KTRs could, in our opinion, become a useful tool for early detection of subclinical processes involving tubules and interstitium (such as acute tubulointerstitial rejection), or timely surveillance of IF/TA. This assumption, however, needs further exploration.

The miRNAs investigated in the present study have so far been extensively related to various physiological and pathological states of kidney graft pathology (see also Supplementary Table S1). We did not observe any association between the circulating levels of selected miRNAs (*miR-29c*, *miR-126*, *miR-146a*, *miR-150*, *miR-155*, and *miR-223*) and kidney graft function (estimated and measured by the reference method at the same time point of miRNA analysis). Previously published studies have shown that the expression of some miRNAs, such as *miR-126* or *miR-223*, is associated with renal function in an up- or downregulated manner [11,15,28]. However, many of those studies were performed with patients with CKD without kidney transplants [15] and/or focused on the functionally unstable period of delayed graft function immediately after transplantation [11,17]. The essential advantage of our analysis is that we used the reference method for GFR measurement (51CrEDTA) and not only routinely used noninvasive biomarkers and/or GFR equations, but the patient cohort in this study was also larger than in most of the so far published research in the field of biomarkers in kidney transplantation [10,11,17,28].

Since selected miRNAs are independent of kidney graft function, they can be reliably used as biomarkers of various pathological processes in KTRs without adjustments to kidney function. In line with previous findings, the results of our pilot study performed on a subgroup of patients

with indications of kidney graft biopsy showed a significant association with the miRNAs selected here, with histologically proven antibody-mediated kidney graft rejection and recurrence of primary glomerulonephritis. In this regard, *miR-29c* expression especially has shown potential for differentiating between these two pathologies (see also Supplementary Figure S1). Unfortunately, the sample examined was too small for reliable inference, but nevertheless pointed to the feasibility of conducting further prospective analysis with systematic serum sampling for miRNA determination and planned implementation of kidney graft biopsy, which is the long term aim of this study.
