First Characterization of ADAMTS-4 in Kidney Tissue and Plasma of Patients with Chronic Kidney Disease—A Potential Novel Diagnostic Indicator

Background: We have previously shown that metzincin protease ADAMTS-4 accompanies renal fibrogenesis, as it appears in the blood of hemodialysis patients. Methods: Native kidney (NKB) and kidney transplant (TXCI) biopsy samples as well as plasma from patients with various stages of CKD were compared to controls. In paired analysis, 15 TXCI samples were compared with their zero-time biopsies (TX0). Tissues were evaluated and scored (interstitial fibrosis and tubular atrophy (IFTA) for NKB and Banff ci for TXCI). Immunohistochemical (IHC) staining for ADAMTS-4 and BMP-1 was performed. Plasma ADAMTS-4 was detected using ELISA. Results: ADAMTS-4 IHC expression was significantly higher in interstitial compartment (INT) of NKB and TXCI group in peritubular capillaries (PTC) and interstitial stroma (INT). Patients with higher stages of interstitial fibrosis (ci > 1 and IFTA > 1) expressed ADAMTS-4 in INT more frequently in both groups (p = 0.005; p = 0.013; respectively). In paired comparison, TXCI samples expressed ADAMTS-4 in INT and PTC more often than TX0. ADAMTS-4 plasma concentration varied significantly across CKD stages, being highest in CKD 2 and 3 compared to other groups (p = 0.0064). Hemodialysis patients had higher concentrations of ADAMTS-4 compared to peritoneal dialysis (p < 0.00001). Conclusion: ADAMTS-4 might have a significant role in CKD as a potential novel diagnostic indicator.


Introduction
Chronic kidney disease (CKD) is a heterogeneous disorder with multiple causes that endangers the structure and function of the kidney. Its occurrence is on a steady rise since the 1990s and has recently reached a global age-standardized prevalence of 9.1% [1]. It is widespread in patients with diabetes, hypertension, obesity and atherosclerosis, affecting around 500 million individuals worldwide [2,3]. CKD is defined by progressive renal deterioration accompanied by loss of function, typically over months or years. Five stages CKD. Expression levels in renal tissue compartments were correlated to CKD progression, and changes of its plasma concentrations were monitored across the same groups.

Sample Collection and Study Outline
This prospective observational study was approved by the Ethics Committee of the University Hospital Center Zagreb (EP-16/106-2). All participants provided a signed informed consent. Blood and tissue samples were collected from patients between January 2018 and June 2019 at the Department of Nephrology, Arterial Hypertension, Dialysis and Transplantation, UHC Zagreb.
Blood samples were drawn into tubes with anticoagulant substance containing 3.8% sodium citrate to form an anticoagulant-to blood ratio (v/v) 1:9. Plasma was obtained by centrifugation (15 min at 4 • C and 3000× g), and aliquots of each sample were stored at −80 • C until analysis. Experimental groups included: 32 patients in CKD stages 1-5, 5 hemodialysis (HD), 7 peritoneal dialysis (PD) patients, 30 kidney transplant patients (TX), and a control group of 5 healthy volunteers (CTRLpl) (Figure 1). concentration of ADAMTS-4 released to the plasma of patients with different stages of CKD. Expression levels in renal tissue compartments were correlated to CKD progression, and changes of its plasma concentrations were monitored across the same groups.

Sample Collection and Study Outline
This prospective observational study was approved by the Ethics Committee of the University Hospital Center Zagreb (EP-16/106-2). All participants provided a signed informed consent. Blood and tissue samples were collected from patients between January 2018 and June 2019 at the Department of Nephrology, Arterial Hypertension, Dialysis and Transplantation, UHC Zagreb.
Blood samples were drawn into tubes with anticoagulant substance containing 3.8% sodium citrate to form an anticoagulant-to blood ratio (v/v) 1:9. Plasma was obtained by centrifugation (15 min at 4 °C and 3000× g), and aliquots of each sample were stored at −80 °C until analysis. Experimental groups included: 32 patients in CKD stages 1-5, 5 hemodialysis (HD), 7 peritoneal dialysis (PD) patients, 30 kidney transplant patients (TX), and a control group of 5 healthy volunteers (CTRLpl) (Figure 1). Kidney biopsy tissue samples were obtained from patients with CKD: 19 patients who underwent native kidney biopsy for clinical diagnosis of kidney disease (NKB group: 9 in CKD stages 1-2 and 10 in CKD stages [3][4][5] and 34 patients at different timepoints after kidney transplantation (TXCI group), who underwent kidney transplant biopsy for clinical evaluation (impairment of kidney function or proteinuria). As controls (CTRL group), Kidney biopsy tissue samples were obtained from patients with CKD: 19 patients who underwent native kidney biopsy for clinical diagnosis of kidney disease (NKB group: 9 in CKD stages 1-2 and 10 in CKD stages [3][4][5] and 34 patients at different timepoints after kidney transplantation (TXCI group), who underwent kidney transplant biopsy for clinical evaluation (impairment of kidney function or proteinuria). As controls (CTRL group), 8 cadaveric kidney donors without evidence of CKD (eGFR > 90 mL/min, absence of proteinuria or any structural changes), who underwent transplant nephrec-tomies, were used ( Figure 1). In 15 subjects from the TXCI group "zero-time", donor kidney biopsy was performed at the day of transplantation (TX0). These samples were obtained from the hospital archive and used for paired analyses (15 TXCI-TX0 pairs) ( Figure 1).
Kidney tissue samples biopsies underwent immunohistochemical (IHC) staining for ADAMTS-4 and BMP-1. Histopathology and immunohistochemistry analyses were performed in Clinical Department for Pathology and Cytology of University Hospital Center Zagreb.
All patient clinical data were obtained from the hospital information system, and clinical data of kidney transplant donors for zero-time kidney biopsies were taken from kidney donor reports.

Kidney Biopsy Sample Collection and Analysis
Native kidney biopsy was performed by standard procedures [18] using ultrasonography and biopsy needle guidance technique. Histologic evaluation by light microscopy and immunofluorescence were performed routinely, combined with electron microscopy in some cases. Interstitial fibrosis and tubular atrophy (IFTA) scores (0-3) were used for the assessment of the chronicity of kidney disease in NKB group. The Banff Working Classification was used for estimation of pathology of kidney transplant biopsy samples (TXCI, TX0) [19].

Immunohistochemistry Analysis (IHC)
Tissue ADAMTS-4 and BMP-1 expression levels were determined on formalinfixed, paraffin embedded (FFPE) tissue sections by IHC. Selected tissues were cut into 4 µm sections, placed on charged slides, and dried at 60 • C for one hour. Slides were cooled to room temperature, deparaffinized with Clarify (American Mastertech, Lodi, CA, USA), and rinsed in water in a Dako Omnis platform (Agilent Technologies, Santa Clara, CA, USA). Epitopes were retrieved with treatment with Dako Flex TRS Low (Agilent Technologies, Santa Clara, CA, USA) for 30 min at 97 • C. Slides were incubated with primary rabbit polyclonal ADAMTS-4 antibody (Abcam, Cambridge, MA, USA) at a dilution 1:200 for 60 min, followed by Dako GV800 HRP-conjugated secondary antibody (Agilent Technologies, Santa Clara, CA, USA) for 20 min. For BMP-1 expression, detection slides were incubated with primary rabbit polyclonal antibody against BMP-1 (HPA014572; Sigma-Aldrich, Merck, Germany) at a dilution 1:50 for 60 min. Thereafter, primary antibody signal amplifier (Rabbit Linker Dako GV809 (Agilent Technologies, Santa Clara, CA, USA)) was applied for 10 min, followed by Dako GV800 HRP-conjugated secondary antibody (Agilent Technologies, Santa Clara, CA, USA) for 20 min. 3,3 -diaminobenzidine chromogen (Agilent Technologies, CA) was applied for 5 min, contrasted with hematoxylin, and cover-slipped. Stained slides were analyzed using an optical microscope (Zeiss Axiostar plus, Oberkochen, Germany; magnification range 200× and 400×). As positive control for ADAMTS-4, middle-sized intralobular artery in kidney biopsy sample was used. ADAMTS-4 was IHC positive in artery wall in the same pattern as previously described by Dong et al. [20].
IHC analyses were performed in three main kidney tissue compartments: (a) interstitial-peritubular capillaries (PTC) and interstitial stroma (INT), (b) glomerular-glomerular capillaries (GC) and Bowman space (BW), (c) tubules-proximal (PXT) and distal (DT). For both molecules (ADAMTS-4 and BMP-1), staining was defined as 0-negative or 1positive for each of the six separate observation areas (PTC, INT, GC, BW, PXT, DT). Positive staining was defined if staining was present in more than 1% of observed area [21]. IHC evaluation was performed independently by two board certified nephropathologists (SB, MC) that were blinded to group allocation.

Enzyme-Linked Immunosorbent Assay (ELISA)
ADAMTS-4 in plasma samples was detected using an indirect ELISA kit (Human ADAMTS-4 DuoSet ELISA DY4307-05 R&D, Minneapolis, MN, USA) according to manufacturer's instructions. Results were obtained with a plate reader (Molecular Devices-SpectraMax i3x) at 450 nm. All samples and standards were analyzed in duplicates and samples with an individual coefficient of variation (CV) greater than 25% were retested in duplicates. All procedures and evaluation of the results were conducted by researchers blinded to clinical and pathological patient data.

Statistical Analysis
Normality of distribution of continuous variables was tested using D'Agostino-Pearson test. Categorical variables were presented as absolute numbers and percentages. Normally distributed continuous variables were presented as mean and standard deviation and nonnormally distributed as median and interquartile range. The frequencies of ADAMTS-4 and BMP-1 proteins in the examined kidney biopsy samples of CTRL, NKB and TXCI group as well as for TX0-TXCI sample pairs were obtained by direct counting based on the positive/negative immunohistochemistry staining. Data distribution differences were compared using Fisher's exact test with Yates correction. Plasma ADAMTS-4 concentration in the group of 8 patients with CKD was presented as multiple variables graphs with box-and-whisker plot for median. The difference in means between the two groups were analyzed by Student's t-test and ANOVA One-Way Analysis of Variance test for more than two independent groups. A two-tailed p-value < 0.05 was used as indicator of statistical significance. All analyses were carried out using MedCalc software (version 19.2.6).

Immunohistochemistry Analysis of ADAMTS-4 and BMP-1 Molecules in Kidney Biopsy Samples of Patients with CKD
Immunohistochemistry analysis included kidney tissue biopsy samples from 19 patients who underwent native kidney biopsy (NKB), 34 patients with transplanted kidney who underwent biopsy due to a clinical indication (TXCI), and 8 cadaveric kidney donors who underwent transplant nephrectomies (CTRL). Study group characteristics are presented in Table 1.

BMP-1 and ADAMTS-4 in Kidney Samples without CKD
BMP-1 was present in proximal tubules of 88% and glomerular Bowman space of 37% subjects in CTRL group and was totally absent in interstitial compartment and glomerular capillaries. ADAMTS-4 was present in tubular (both in PXT and DT) and glomerular (both in GC and BW) compartments of samples without CKD. Only 13% of subjects expressed ADAMTS-4 in interstitial stroma and none in peritubular capillaries.
ADAMTS-4 and BMP-1 IHC expression in kidney biopsy samples from NKB group is presented in Figure 3.

Analysis of ADAMTS-4 IHC Expression in 15 Transplant Kidney Biopsy Sample Pairs (TX0-TXCI)
The results of a paired sub-analysis comparing 15 kidney biopsy samples from TXCI group performed at different timepoints after transplantation with their paired (same allograft) "zero time "samples (TX0) performed at the day of transplantation (15 TX0-TXCI pairs) are presented in Table 2. In paired analysis, subjects in TXCI group had significantly higher CKD stage and ci score than in TX0 group (TXCI vs. TX0, all p < 0.001, Table 2 ADAMTS-4 expression by IHC staining for one pair of TXCI-TX0 biopsy samples is presented in Figure 4.

Detection of ADAMTS-4 in Plasma of Patients with CKD (Elisa Analysis)
Distribution of ADAMTS-4 concentration in plasma of the group of eight patients with CKD (CKD stages 1-5, PD, HD before and after procedure: HD-B and HD-A, TX) and controls (CTRLpl) is presented in Figure 5. When comparing mean ADAMTS-4 plasma concentrations between CTRLpl and all studied groups (CKD 1-5, PD, HD, TX), significant difference was found between CTRLpl and HD group (CTRL vs. HD-B; 226 vs. 1032.6; p = 0.006). ADAMTS-4 plasma concentration varied significantly between CKD stages, being highest in CKD stages 2 and 3, taken together and compared to all other CKD stages (CKD 2 + 3 vs. 1 + 4 + 5; 2925.1 vs. 613.6; respectively p = 0.0064) or to all studied groups taken together (CKD 2 + 3 vs. CKD1 + 4 + 5 + HD-B + PD + TX; p = 0.00004) Interestingly, when analyzing only dialysis groups, none of the patients in PD group had detectable ADAMTS-4 in plasma, which differed significantly from the HD group (PD vs. HD-B; 0 vs. 1032.6, respectively; p < 0.00001). Additionally, ADAMTS-4 plasma values in HD group were not different even after a regular 4 h hemodialysis procedure (HD-B vs. HD-A; 1032.6 vs. 1171, respectively; p = 0.084).

Detection of ADAMTS-4 in Plasma of Patients with CKD (Elisa Analysis)
Distribution of ADAMTS-4 concentration in plasma of the group of eight patients with CKD (CKD stages 1-5, PD, HD before and after procedure: HD-B and HD-A, TX) and controls (CTRLpl) is presented in Figure 5. When comparing mean ADAMTS-4 plasma concentrations between CTRLpl and all studied groups (CKD 1-5, PD, HD, TX), significant difference was found between CTRLpl and HD group (CTRL vs. HD-B; 226 vs. 1032.6; p = 0.006). ADAMTS-4 plasma concentration varied significantly between CKD stages, being highest in CKD stages 2 and 3, taken together and compared to all other CKD stages (CKD 2 + 3 vs. 1 + 4 + 5; 2925.1 vs. 613.6; respectively p = 0.0064) or to all studied groups taken together (CKD 2 + 3 vs. CKD1 + 4 + 5 + HD-B + PD + TX; p = 0.00004) Interestingly, when analyzing only dialysis groups, none of the patients in PD group had detectable ADAMTS-4 in plasma, which differed significantly from the HD group (PD vs. HD-B; 0 vs. 1032.6, respectively; p < 0.00001). Additionally, ADAMTS-4 plasma values in HD group were not different even after a regular 4 h hemodialysis procedure (HD-B vs. HD-A; 1032.6 vs. 1171, respectively; p = 0.084). Significant difference of mean ADAMTS-4 plasma concentrations was found between CTRLpl and HD-B group (p = 0.006) and also varied significantly between CKD stages being highest in patients with CKD stages 2 and 3 compared to CKD 1 + 4 + 5 (p = 0.0064) and CKD 1 + 4 + 5 + HD-B + PD + TX (p = 0.00004). None of the patients in PD group had detectable plasma ADAMTS-4, which differed significantly from the HD group (p < 0.00001).

Discussion
In this study, to the best of our knowledge, for the first time, ADAMTS-4 was characterized in kidney tissues and plasma of patients with CKD. Previous research of ADAMTS-4 in kidney diseases is scarce. Wang et al. reported an elevated ADAMTS-4 expression in patients with early stage membranous nephropathy; however, because of the use of tissue lysates, the exact origin of the protein could not be determined [22]. A single animal study indicated that ADAMTS-4 expression in healthy murine kidneys occurs in proximal and distal convoluted tubules [23]. We have determined that patients of the native kidney biopsy group (NKB) with early-stage CKD predominantly express ADAMTS-4 in the glomerular and tubular and only slightly in the interstitial compartment. Taking into account that almost 90% of our NKB group was diagnosed with primary glomerular disease, our results concur with the Wang group.
The underlying pathology of CKD and ESRD is renal fibrosis, which initiates mainly in the peritubular capillaries of the interstitial compartment [9,24]. Other ADAMTS proteins have previously been implicated in the pathogenesis of kidney fibrosis: higher expression of ADAMTS-1 was detected in the early course of fibrogenesis on mouse model of unilateral ureteral obstruction. Investigators defined endothelial cell of PTC as the source of ADAMTS-1 in the process of pericytes detachment and PTC destabilization [25]. Significant difference of mean ADAMTS-4 plasma concentrations was found between CTRLpl and HD-B group (p = 0.006) and also varied significantly between CKD stages being highest in patients with CKD stages 2 and 3 compared to CKD 1 + 4 + 5 (p = 0.0064) and CKD 1 + 4 + 5 + HD-B + PD + TX (p = 0.00004). None of the patients in PD group had detectable plasma ADAMTS-4, which differed significantly from the HD group (p < 0.00001).

Discussion
In this study, to the best of our knowledge, for the first time, ADAMTS-4 was characterized in kidney tissues and plasma of patients with CKD. Previous research of ADAMTS-4 in kidney diseases is scarce. Wang et al. reported an elevated ADAMTS-4 expression in patients with early stage membranous nephropathy; however, because of the use of tissue lysates, the exact origin of the protein could not be determined [22]. A single animal study indicated that ADAMTS-4 expression in healthy murine kidneys occurs in proximal and distal convoluted tubules [23]. We have determined that patients of the native kidney biopsy group (NKB) with early-stage CKD predominantly express ADAMTS-4 in the glomerular and tubular and only slightly in the interstitial compartment. Taking into account that almost 90% of our NKB group was diagnosed with primary glomerular disease, our results concur with the Wang group.
The underlying pathology of CKD and ESRD is renal fibrosis, which initiates mainly in the peritubular capillaries of the interstitial compartment [9,24]. Other ADAMTS proteins have previously been implicated in the pathogenesis of kidney fibrosis: higher expression of ADAMTS-1 was detected in the early course of fibrogenesis on mouse model of unilateral ureteral obstruction. Investigators defined endothelial cell of PTC as the source of ADAMTS-1 in the process of pericytes detachment and PTC destabilization [25]. ADAMTS-2 and -12 were upregulated in mesenchymal interstitial cells during their transformation to myofibroblasts [26]. Furthermore, ADAMTS-1, -12, and -15 were found upregulated in animal model of Adriamycin-induced kidney fibrosis [27]. Our results implicate ADAMTS-4 in the progression of CKD and thus possibly the pathogenesis of kidney fibrosis. Indeed, there is no clear evidence on the mechanism of action of ADAMTS-4 in renal failure although its function may be assumed by studying its role in other organ systems. Boyd et al. showed that fibroblasts produce ADAMTS-4 in response to tissue-damage, which leads to ECM remodeling and promotion of immune cell infiltration in the lungs [28]. Additionally, the circulating active form of the protein may play a role in the regulation of ECM expression and deposition, ultimately leading to fibrosis and a disease state. ADAMTS-4 expression in the interstitial compartment was elevated across all disease groups; however, it was most abundantly expressed in non-transplanted patients with advanced disease stages (CKD 3-5). Similar localization of ADAMTS-4 in the interstitial compartment among equivalent CKD stages of NKB and TXCI groups implicates a similar role of ADAMTS-4 in CKD progressivity. Furthermore, increased expression of ADAMTS-4 in the interstitial compartment significantly correlated to higher chronic interstitial fibrosis scores (IFTA for NKB and ci for TXCI). Finally, paired consecutive biopsy samples from the transplant group clearly confirmed that ADAMTS-4 expression rises proportionally to the amount of interstitial fibrosis (ci score) and stage of CKD. ADAMTS-4 expression was also detected in distal tubules of almost all biopsy samples regardless of presence of CKD and to significantly lesser extent in proximal tubules, whereas its expression was significantly associated with the presence of acute tubular injury.
These effects might be partially explained by the role ADAMTS-4 has in the processing of extracellular matrix [29]. Expression of versican, a key ADAMTS-4 substrate has been associated to CKD progression in animal models and confirmed in patients with proteinuric kidney diseases [14,30,31]. Inflammatory cells, namely macrophages, seem to be important in the early phase of CKD and kidney fibrosis [26,32,33]. This pro-inflammatory milieu drives the expression of ADAMTS-4 enzymes by macrophages and macrophage-like cells derived from vascular smooth muscle cells. A positive feedback loop is supported where its substrates versican and aggrecan enhance plaque formation, thus fueling fibrosis. The cleavage of these proteoglycans by ADAMTS-4 in turn further destabilizes the atherosclerotic plaques, leading to unfavorable patient outcomes [34,35]. The mechanisms of action of ADAMTS-4 proposed in atherosclerosis might also be relevant in CKD; however, this remains to be elucidated [15,17,20]. We have shown that ADAMTS-4 expression is enhanced in the glomerular capillaries of patients with CKD for both study groups (NKB and TXCI) as compared to controls. The significance of such correlation is yet to be explored, but presumably, ADAMTS-4 might be expressed by activated circulating macrophages as a part of kidney disease. To further assess the role of ADAMTS4 in CKD fibrosis, we also tested the expression of BMP-1, a tolloid-like proteinase, originally identified for its roles in the maturation of procollagen [36]. Both molecules, BMP-1 and ADAMTS-4, were detected in different tubular compartments. Although previous studies found that early inhibition of BMP1-3 significantly reduced kidney fibrosis in rats [12,37], we saw no correlation of BMP-1 expression to CKD progression.
The analysis of plasma ADAMTS-4 revealed its expression across all stages of CKD, however, with significant variations. As the disease progresses, the expression levels peak at CKD stages 2 and 3 and afterwards sharply decrease towards ESRD. It could be speculated that higher systemic concentrations of ADAMTS-4 might be the result of inflammation that underlies kidney diseases. This is especially the case in primary glomerular disease; however, larger studies are needed for any further conclusions. In renal replacement therapies, ADAMTS-4 was detected in patients on hemodialysis before but also after dialysis. This finding is in concordance with previous research and could be explained by the fact that ADAMTS-4 is a large uremic toxin, which is poorly filtrated using a standard dialysis filter [13]. Interestingly, none of the patients on peritoneal dialysis had detectable plasma ADAMTS-4. As peritoneal dialysis has different capacity for large molecule excretion, it can be assumed that ADAMTS-4 could be excreted using this treatment, but this should be additionally tested. Alternatively, this discrepancy could stem from the fact that hemodialysis patients tend to have more comorbidities, especially cardiovascular diseases and atherosclerosis, which could be independent sources of elevated plasma ADAMTS-4 [38]. That being said, the main limitation of our study is also its relatively small sample size. The presented concepts should be tested further on larger pool of patients with stratified primary kidney disease and various stages of CKD. For similar reasons, as the characterization of ADAMTS-4 in kidney samples from CKD patients was done for the first time by IHC, we opted not to apply intensification of IHC staining, which could be used in further studies.

Conclusions
This study characterized ADAMTS-4 proteinases for the first time in kidney tissue samples and plasma of patients with various stages of CKD. Based on its significantly higher IHC expression in peritubular capillaries and interstitial space of patients with CKD compared to controls, its stepwise abundance along with progression of interstitial fibrosis, which was also confirmed on the paired biopsy samples of the 15 transplanted kidneys, ADAMTS-4 might be a novel marker of CKD. Our preliminary results of plasma analysis also confirmed the presence of circulating ADAMTS-4 in CKD patients, being highest in stages 2 and 3, with subsequent disappearance until stage 5, which can fit into the dynamics of the development of the disease. Further prospective studies in larger cohorts are needed to validate the results of the present study. Funding: Scientific Center of Excellence for Reproductive and Regenerative Medicine (project "Reproductive and regenerative medicine-exploration of new platforms and potentials", Grant Agreement KK.01.1.1.01.0008, which is funded by the European Union through the European Regional Development Fund.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of University Hospital Center Zagreb (EP-16/106-2).
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.