Background. An ELISA to analyse uromodulin in human serum (sUmod) was developed, validated and tested for clinical applications.

Methods. We assessed sUmod, a very stable antigen, in controls, patients with chronic kidney disease (CKD) stages 1–5, persons with autoimmune kidney diseases and recipients of a renal allograft by ELISA.

Results. Median sUmod in 190 blood donors was 207 ng/mL (women: men, median 230 versus 188 ng/mL, P = 0.006). sUmod levels in 443 children were 193 ng/mL (median). sUmod was correlated with cystatin C (rs = −0.862), creatinine (rs = −0.802), blood urea nitrogen (BUN) (rs = −0.645) and estimated glomerular filtration rate (eGFR)–cystatin C (rs  = 0.862). sUmod was lower in systemic lupus erythematosus-nephritis (median 101 ng/mL), phospholipase-A2 receptor- positive glomerulonephritis (median 83 ng/mL) and anti-glomerular basement membrane positive pulmorenal syndromes (median 37 ng/mL). Declining sUmod concentrations paralleled the loss of kidney function in 165 patients with CKD stages 1–5 with prominent changes in sUmod within the ‘creatinine blind range’ (71–106 µmol/L). Receiver-operating characteristic analysis between non-CKD and CKD-1 was superior for sUmod (AUC 0.90) compared with eGFR (AUC 0.39), cystatin C (AUC 0.39) and creatinine (AUC 0.27). sUmod rapidly recovered from 0 to 62 ng/mL (median) after renal transplantation in cases with immediate graft function and remained low in delayed graft function (21 ng/mL, median; day 5–9: relative risk 1.5–2.9, odds ratio 1.5–6.4). Immunogold labelling disclosed that Umod is transferred within cytoplasmic vesicles to both the apical and basolateral plasma membrane. Umod revealed a disturbed intracellular location in kidney injury.

Conclusions. We conclude that sUmod is a novel sensitive kidney-specific biomarker linked to the structural integrity of the distal nephron and to renal function.

INTRODUCTION

Uromodulin (Umod) is a glycoprotein exclusively synthesized in the human kidney by epithelial cells of the thick ascending limb (TAL) of the loop of Henle and the early distal convoluted tubule [1–4]. This glycosylphosphatidylinositol (GPI)-anchored protein is predominantly localized to the apical plasma membrane, from which it is released into the tubular lumen by proteolytic cleavage [5]. Umod of urinary origin (uUmod) is a high molecular weight polymer and the most abundant urinary protein, detectable at a concentration of approximately 75 mg/24 h [6–8]. As a renal defensin, Umod protects against urinary tract infection by clotting fibrillary type-1 and type-S adhesins of bacteria [6–8]. Furthermore, uUmod prevents renal stone formation [9], is involved in handling tubular Na+/K+/Cl transport and is suspected to contribute to arterial hypertension [10–12]. Umod exhibits immunomodulatory properties, activates granulocytes, mediates monocyte cytotoxicity and inhibits T-cell proliferation [13–15]. In vitro studies showed that Umod can be either pro- or anti-inflammatory [16, 17]. Epithelia of the TAL are specifically susceptible to oxidative stress, hypoxia and nephrotoxins. There, vulnerable, sublethally damaged cells may undergo apoptosis and necrosis, finally resulting in acute kidney injury (AKI) as was shown in animal models [18–23].

Up to now very little attention has been paid to the fact that uromodulin is also present in serum (sUmod; [24–26]). This might originate from a frequently overlooked basolateral release of Umod into the interstitium that was shown in a mouse model [27].

Various studies have focused on the role of uUmod in chronic kidney diseases (CKD) of different a aetiologies [28–31]. Excretion of uUmod was reduced in patients with CKD or chronic graft failure and paralleled the estimated glomerular filtration rate (eGFR) [32–34]. In contrast, a case–control study reported that increased uUmod concentrations precede the development of CKD [35]. UMOD was confirmed as one of the most important loci associated with end-stage renal failure [36–42].

Preanalytic data showed that the urinary form impairs precision and reliability of the measurements [2, 31, 43]. It is speculated that the polymeric structure and varying degrees of polymerization hamper antibody-based detection.

In the present work, we describe the development and application of an ELISA for the determination of sUmod, based on monoclonal antibodies (moAb). We assessed sUmod levels in healthy persons, patients with different stages of CKD and patients after kidney transplantation (KTX). Our immunohistochemical findings support the view that newly biosynthesized Umod is translocated and released bidirectionally by TAL cells, such that the basolateral pathway might contribute to sUmod levels. The observed progressive decrease of sUmod levels related to CKD stages suggests altered cell viability in the TAL segment, which mirrors a gradual and more global renal parenchymal injury.

MATERIALS AND METHODS

Study populations

Physiological levels of sUmod were assessed in 190 blood donors (18–60 years, Table 1) and in 443 children and adolescents (1–17 years). Thirty-three patients without kidney disease (non-CKD) and 132 patients with CKD stages 1–5 were enrolled in a nonrandomized manner. This cohort was used to determine the correlation of sUmod with cystatin C, creatinine and eGFR (Table 2). sUmod was studied in patients with autoimmune diseases classified by clinical symptoms, CKD stages and autoantibody tests (Euroimmun, Lübeck, Germany), i.e. systemic lupus erythematosus (SLE; n = 53), anti-glomerular basement membrane (GBM)-positive pulmorenal syndromes (n = 10), phospholipase-A2 receptor (PLA2-R) antibody (Ab)-positive glomerulonephritis (n = 20). Forty-four kidney transplant patients [17 with delayed graft function (DGF) and 27 with immediate graft function (IGF)] were analysed for sUmod levels (Table 3). Details of immunosuppressive therapy in transplant patients have been published elsewhere [44]. All individuals underwent physical examination and completed a detailed questionnaire. Ethics approval was obtained from two ethics committees and all participants provided informed consent, in adherence to the declarations of Helsinki and Istanbul.

Table 1

Demographic characteristics and serum uromodulin concentration of blood donors; data are presented as mean ± SD and median, with IQR in brackets

Parameter Total (n = 190) Male (n = 89) Female (n = 101) P-value 
Age (years) 38.7 ± 11.9, 40.0 (28.0, 49.0) 37.4 ± 11.8, 38.0 (25.5, 47.5) 39.8 ± 11.9, 42.0 (29.8, 50) 0.165 
Uromodulin (ng/mL) 221.5 ± 95.2, 207.4 (148.5, 275.1) 199.3 ± 78.5, 188.2 (144.4, 250.7) 241.1 ± 103.5, 229.9 (164.7, 307.6) 0.006 
Parameter Total (n = 190) Male (n = 89) Female (n = 101) P-value 
Age (years) 38.7 ± 11.9, 40.0 (28.0, 49.0) 37.4 ± 11.8, 38.0 (25.5, 47.5) 39.8 ± 11.9, 42.0 (29.8, 50) 0.165 
Uromodulin (ng/mL) 221.5 ± 95.2, 207.4 (148.5, 275.1) 199.3 ± 78.5, 188.2 (144.4, 250.7) 241.1 ± 103.5, 229.9 (164.7, 307.6) 0.006 
Table 2

Demographic characteristics and laboratory blood parameters; data are presented as mean ± SD and median, with IQR in brackets, or as absolute numbers (n) with percentage (% of the referring group) in brackets

 Total (n = 165) Non-CKD (n = 33, 20.0%) CKD 1 (n = 21, 12.7%) CKD 2 (n = 27, 16.4%) CKD 3 (n = 43, 26.1%) CKD 4 (n = 38, 23.0%) CKD 5 (n = 3, 1.8%) P-value 
Age (years) 61.9 ± 14.4, 65.0 (51.0, 73.0) 60.8 ± 14.2, 66.0 (62.0, 67.0) 56.0± 11.3, 55.0 (45.0, 63.0) 60.5 ± 12.4, 62.0 (48.0, 72.0) 64.0 ± 17.6, 71.0 (52.5, 77.0) 63.9 ± 12.1, 66.0 (56.3, 73.0) 63.3 ± 16.0, 53.0 (52.0, 69.5) 0.159 
Gender (m/f in %) 110/55 (66.7/33.3) 20/13 (60.6/39.4) 15/6 (71.4/28.6) 18/9 (66.7/33.3) 28/15 (65.1/34.9) 26/12 (68.4/31.6) 3/0 (100/0) 0.888 
Hypertension 74 (44, 8%) 1 (3%) 12 (57%) 17 (60, 7%) 19 (44%) 23 (60, 5%)  
NIDDM 27 6 (28, 5%) 5 (17, 8%) 8 (18, 6%) 7 (18, 4%)  
IDDM 13 5 (11, 6%) 8 (21%)  
GN, others  1 former GN in complete remission mGN (1) SLE (1) ANCA (1) GN 7 (25%), pHPT (2), sarkoidosis (3), myeloma (2) GN 7 (16, 2%), PN 2 (4, 6%), myeloma 6 (13, 9%) GN 8 (21%), toxN 6 (15, 7%), obstrN 3 (7, 9%) mGN (1 
Proteinuria <1 g/g creat  9 (42%) 13 (46%) 12 (28%) 15 (39%)  
Proteinuria >1 g/g creat  1 (4, 7%) 2 (7, 1%) 8 (18, 6%) 15 (31, 8)  
Uric acid (µmol/L) 375.3 ± 93.9, 369.0 (315.5, 424.1) 316.0 ± 79.7, 315.5 (267.9, 381.0) 353.4 ± 81.0, 330.4 (293.2, 431.5) 332.5 ± 76.6, 333.3 (273.8, 378.0) 391.9 ± 64.2, 372.0 (345.2, 422.6) 469.0 ± 92.8, 422.6 (401.8, 523.8) 333.3 ± 0.0, 333.3 (333.3, 333.3) <0.001 
C-reactive protein (mg/dL) 1.3 ± 1.9, 0.9 (0.4, 1.5) 0.6 ± 0.2, 0.5 (0.4, 0.6) 0.8 ± 0.5, 0.7 (0.6, 0.9) 1.4 ± 0.2, 1.4 (1.3, 1.5) 0.7 ± 0.4, 0.7 (0.4, 1.1) 2.5 ± 3.2, 1.7 (1.2, 2.0) 1.4 ± 0.0, 1.4 (1.4, 1.4) <0.001 
Urea (mmol/L) 11.9 ± 8.0, 8.4 (5.9, 16.7) 5.9 ± 1.6, 5.7 (4.9, 6.8) 5.1 ± 1.2, 4.9 (4.6, 5.9) 7.3 ± 3.4, 6.2 (6.0, 7.5) 11.9 ± 4.4, 11.8 (7.9, 15.5) 21.8 ± 6.5, 20.2 (16.7, 26.7) 28.4 ± 4.3, 29.5 (26.1, 31.2) <0.001 
Sodium (mmol/L) 138.0 ± 12.1, 139.0 (138.0, 141.0) 135.1 ± 19.9, 139.0 (138.0, 140.0) 138.2 ± 3.8, 138.0 (137.0, 141.0) 139.4 ± 2.9, 139.0 (139.0, 140.0) 139.1 ± 4.0, 139.0 (137.5, 141.0) 137.4 ± 17.0 140.0, (139.0, 142.8) 141.7 ± 0.5, 142.0 (141.5, 142.0) 0.088 
Potassium (mmol/L) 4.4 ± 0.6, 4.3 (4.0, 4.7) 4.2 ± 0.4, 4.1 (3.9, 4.5) 4.2 ± 0.9, 4.0 (3.9, 4.3) 4.3 ± 0.5, 4.3 (4.0, 4.6) 4.5 ± 0.5, 4.4 (4.2, 4.8) 4.7 ± 0.7, 4.7 (4.2, 5.1) 4.7 ± 0.2, 4.7 (4.6, 4.9) <0.001 
Calcium (mmol/L) 2.3 ± 0.2, 2.4 (2.3, 2.4) 2.4 ± 0.1, 2.4 (2.3, 2.5) 2.4 ± 0.1, 2.4 (2.3, 2.4) 2.4 ± 0.1, 2.4 (2.3, 2.4) 2.3 ± 0.2, 2.3 (2.3, 2.4) 2.36 ± 0.2, 2.3 (2.2, 2.4) 2.3 ± 0.1, 2.3 (2.3, 2.4) 0.096 
Phosphate (mmol/L) 1.2 ± 0.3, 1.1 (1.0, 1.3) 1.1 ± 0.2, 1.1 (1.0, 1.2) 1.0 ± 0.2, 1.1 (0.9, 1.1) 1.0 ± 0.2, 1.0 (0.9, 1.2) 1.1 ± 0.2, 1.1 (0.9, 1.3) 1.4 ± 0.5, 1.3 (1.1, 1.5) 1.6 ± 0.6, 1.6 (1.3, 2.0) <0.001 
Total protein (g/L) 75 ± 28, 73 (67, 76) 73 ± 5, 73 (69, 77) 73 ± 6, 74 (67, 77) 73 ± 6, 74 (69, 76) 74 ± 10, 73 (70, 76) 84 ± 63, 67 (65, 73) 73 ± 0, 73 (73, 73) 0.325 
Cholesterol (mmol/L) 5.7 ± 1.6, 5.8 (4.6, 7.0) 6.6 ± 1.0, 6.6 (5.8, 7.4) 6.0 ± 1.3, 6.1 (5.3, 7.2) 5.5 ± 1.3, 5.1 (4.8, 6.4) 4.8 ± 1.8, 4.4 (3.7, 5.0) 6.0 ± 1.3, 5.9 (5.1, 6.9) 4.9 ± 0.0, 4.9 (4.9 , 4.9) <0.001 
HDL (mmol/L) 1.5 ± 0.4, 1.4 (1.2, 1.7) 1.6 ± 0.3, 1.6 (1.4, 1.8) 1.4 ± 0.5, 1.3 (1.0, 1.7) 1.4 ± 0.5, 1.2 (1.1, 1.7) 1.4 ± 0.4, 1.4 (1.2, 1.6) 1.5 ± 0.6, 1.3 (1.2, 1.6) 1.1 ± 0.0, 1.1 (1.1, 1.1) 0.112 
LDL (mmol/L) 4.0 ± 1.3, 4.0 (3.0, 5.0) 4.6 ± 0.8, 4.6 (4.0, 5.1) 4.4 ± 1.2, 4.4 (3.9, 5.0) 3.7 ± 0.8, 3.7 (3.1, 4.5) 3.2 ± 1.4, 2.7 (2.4, 3.2) 4.0 ± 1.3, 4.1 (3.1, 4.9) 3.6 ± 0.0, 3.6 (3.6, 3.6) <0.001 
Triglyceride (mmol/L) 1.9 ± 1.4, 1.5 (1.0, 2.2) 1.5 ± 0.5, 1.5 (1.1, 1.9) 1.8 ± 0.8, 1.7 (1.3, 1.9) 2.4 ± 2.1, 1.7 (1.1, 2.6) 1.7 ± 1.2, 1.2 (0.9, 2.2) 2.4 ± 1.7, 2.1 (1.1, 3.0) 1.9 ± 0.0, 1.9 (1.9, 1.9) 0.256 
Haemoglobin (mmol/L) 8.3 ± 1.6, 8.7 (7.9, 9.2) 8.9 ± 0.5, 8.9 (8.4, 9.2) 8.8 ± 1.5, 9.2 (8.7, 9.5) 8.2 ± 1.8, 8.6 (7.9, 9.2) 8.0 ± 2.1, 8.1 (7.0, 9.0) 7.7 ± 1.4, 8.2 (6.8, 8.9) 7.8 ± 0.0, 7.8 (7.8, 7.8) 0.006 
Leucocyte count (10e9/L) 8.0 ± 7.4, 6.9 (5.9, 8.2) 8.1 ± 11.1, 6.4 (5.4, 7.1) 10.0 ± 10.0, 6.6 (6.1, 9.5) 7.6 ± 2.6, 8.0 (6.8, 8.6) 7.5 ± 4.6, 6.9 (6.1, 7.8) 7.2 ± 1.3, 7.4 (6.2, 8.4) 9.2 ± 0.0, 9.2 (9.2, 9.2) 0.057 
Kidney biomarker level 
Uromodulin (ng/mL) 115.6 ± 101.0, 84.8 (37.0, 180.0) 260.4 ± 101.6, 227.6 (194.4, 277.8) 150.7 ± 54.8, 153.4 (121.5, 180.7) 116.7 ± 44.5, 111.8 (80.9, 139.4) 70.5 ± 47.1, 51.8 (36.1, 86.2) 29.1 ± 13.2, 30.3 (19.2, 38.0) 10.3 ± 7.0, 12.7 (6.7, 15.1) <0.001 
Cystatin C (nmol/L) 147.1 ± 93.8, 113.8 (72.7, 206.3) 66.7 ± 10.1, 68.2 (62.9, 73.4) 62.1 ± 10.9, 66.7 (55.4, 70.4) 91.5 ± 9.5, 89.1 (85.0, 98.9) 151.9 ± 27.4, 147.6 (127.0, 176.4) 273.8 ± 39.3, 271.1 (247.9, 304.3) 453.1 ± 70.5, 445.7 (408.2, 494.3) <0.001 
Creatinine (µmol/L) 176.9 ± 126.1, 127.0 (93.0, 223.3) 96.7 ± 22.1, 94.0 (80.0, 117.0) 82.4 ± 16.8, 76.9 (72.0, 93.0) 107.0 ± 21.8, 105.0 (91.5, 119.2) 169.3 ± 52.7, 160.9 (125.0, 206.0) 333.5 ± 131.5, 316.5 (246.1, 383.1) 669.5 ± 153.0, 670.0 (575.9, 763.3) <0.001 
BUN (mmol/L) 11.9 ± 8.0, 8.4 (5.9, 16.6) 5.9 ± 1.6, 5.7 (4.9, 6.8) 5.1 ± 1.2, 4.9 (4.6 , 5.9) 7.3 ± 3.4, 6.2 (6.0 , 7.4) 11.9 ± 4.4, 11.8 (7.9 , 15.5) 21.8 ± 6.5, 20.2 (16.7 , 26.7) 28.3 ± 4.3, 29.4 (26.0 , 31.2) <0.001 
eGFRHoek (mL/min) 64.2 ± 37.9, 57.7 (29.7, 92.9) 104.6 ± 20.7, 99.3 (91.9, 108.0) 114.0 ± 26.8, 101.0 (96.0, 123.2) 73.6 ± 7.8, 74.9 (67.1, 78.8) 43.7 ± 8.7, 43.6 (35.7, 51.1) 22.0 ± 3.8, 21.8 (19.1, 24.3) 11.8 ± 2.5, 12.0 (10.3, 13.4) <0.001 
eGFRCKD-EPI (mL/min) 49.4 ± 31.1, 44.0 (20.5, 71.5) 78.8 ± 17.5, 76.5 (65.1, 86.2) 92.8 ± 18.3, 88.9 (85.1, 97.0) 60.1 ± 11.5, 61.6 (51.5, 69.5) 32.5 ± 8.7, 31.4 (25.7, 39.5) 14.5 ± 4.0, 14.2 (11.6, 16.9) 6.7 ± 1.2, 7.0 (6.0, 7.5) <0.001 
 Total (n = 165) Non-CKD (n = 33, 20.0%) CKD 1 (n = 21, 12.7%) CKD 2 (n = 27, 16.4%) CKD 3 (n = 43, 26.1%) CKD 4 (n = 38, 23.0%) CKD 5 (n = 3, 1.8%) P-value 
Age (years) 61.9 ± 14.4, 65.0 (51.0, 73.0) 60.8 ± 14.2, 66.0 (62.0, 67.0) 56.0± 11.3, 55.0 (45.0, 63.0) 60.5 ± 12.4, 62.0 (48.0, 72.0) 64.0 ± 17.6, 71.0 (52.5, 77.0) 63.9 ± 12.1, 66.0 (56.3, 73.0) 63.3 ± 16.0, 53.0 (52.0, 69.5) 0.159 
Gender (m/f in %) 110/55 (66.7/33.3) 20/13 (60.6/39.4) 15/6 (71.4/28.6) 18/9 (66.7/33.3) 28/15 (65.1/34.9) 26/12 (68.4/31.6) 3/0 (100/0) 0.888 
Hypertension 74 (44, 8%) 1 (3%) 12 (57%) 17 (60, 7%) 19 (44%) 23 (60, 5%)  
NIDDM 27 6 (28, 5%) 5 (17, 8%) 8 (18, 6%) 7 (18, 4%)  
IDDM 13 5 (11, 6%) 8 (21%)  
GN, others  1 former GN in complete remission mGN (1) SLE (1) ANCA (1) GN 7 (25%), pHPT (2), sarkoidosis (3), myeloma (2) GN 7 (16, 2%), PN 2 (4, 6%), myeloma 6 (13, 9%) GN 8 (21%), toxN 6 (15, 7%), obstrN 3 (7, 9%) mGN (1 
Proteinuria <1 g/g creat  9 (42%) 13 (46%) 12 (28%) 15 (39%)  
Proteinuria >1 g/g creat  1 (4, 7%) 2 (7, 1%) 8 (18, 6%) 15 (31, 8)  
Uric acid (µmol/L) 375.3 ± 93.9, 369.0 (315.5, 424.1) 316.0 ± 79.7, 315.5 (267.9, 381.0) 353.4 ± 81.0, 330.4 (293.2, 431.5) 332.5 ± 76.6, 333.3 (273.8, 378.0) 391.9 ± 64.2, 372.0 (345.2, 422.6) 469.0 ± 92.8, 422.6 (401.8, 523.8) 333.3 ± 0.0, 333.3 (333.3, 333.3) <0.001 
C-reactive protein (mg/dL) 1.3 ± 1.9, 0.9 (0.4, 1.5) 0.6 ± 0.2, 0.5 (0.4, 0.6) 0.8 ± 0.5, 0.7 (0.6, 0.9) 1.4 ± 0.2, 1.4 (1.3, 1.5) 0.7 ± 0.4, 0.7 (0.4, 1.1) 2.5 ± 3.2, 1.7 (1.2, 2.0) 1.4 ± 0.0, 1.4 (1.4, 1.4) <0.001 
Urea (mmol/L) 11.9 ± 8.0, 8.4 (5.9, 16.7) 5.9 ± 1.6, 5.7 (4.9, 6.8) 5.1 ± 1.2, 4.9 (4.6, 5.9) 7.3 ± 3.4, 6.2 (6.0, 7.5) 11.9 ± 4.4, 11.8 (7.9, 15.5) 21.8 ± 6.5, 20.2 (16.7, 26.7) 28.4 ± 4.3, 29.5 (26.1, 31.2) <0.001 
Sodium (mmol/L) 138.0 ± 12.1, 139.0 (138.0, 141.0) 135.1 ± 19.9, 139.0 (138.0, 140.0) 138.2 ± 3.8, 138.0 (137.0, 141.0) 139.4 ± 2.9, 139.0 (139.0, 140.0) 139.1 ± 4.0, 139.0 (137.5, 141.0) 137.4 ± 17.0 140.0, (139.0, 142.8) 141.7 ± 0.5, 142.0 (141.5, 142.0) 0.088 
Potassium (mmol/L) 4.4 ± 0.6, 4.3 (4.0, 4.7) 4.2 ± 0.4, 4.1 (3.9, 4.5) 4.2 ± 0.9, 4.0 (3.9, 4.3) 4.3 ± 0.5, 4.3 (4.0, 4.6) 4.5 ± 0.5, 4.4 (4.2, 4.8) 4.7 ± 0.7, 4.7 (4.2, 5.1) 4.7 ± 0.2, 4.7 (4.6, 4.9) <0.001 
Calcium (mmol/L) 2.3 ± 0.2, 2.4 (2.3, 2.4) 2.4 ± 0.1, 2.4 (2.3, 2.5) 2.4 ± 0.1, 2.4 (2.3, 2.4) 2.4 ± 0.1, 2.4 (2.3, 2.4) 2.3 ± 0.2, 2.3 (2.3, 2.4) 2.36 ± 0.2, 2.3 (2.2, 2.4) 2.3 ± 0.1, 2.3 (2.3, 2.4) 0.096 
Phosphate (mmol/L) 1.2 ± 0.3, 1.1 (1.0, 1.3) 1.1 ± 0.2, 1.1 (1.0, 1.2) 1.0 ± 0.2, 1.1 (0.9, 1.1) 1.0 ± 0.2, 1.0 (0.9, 1.2) 1.1 ± 0.2, 1.1 (0.9, 1.3) 1.4 ± 0.5, 1.3 (1.1, 1.5) 1.6 ± 0.6, 1.6 (1.3, 2.0) <0.001 
Total protein (g/L) 75 ± 28, 73 (67, 76) 73 ± 5, 73 (69, 77) 73 ± 6, 74 (67, 77) 73 ± 6, 74 (69, 76) 74 ± 10, 73 (70, 76) 84 ± 63, 67 (65, 73) 73 ± 0, 73 (73, 73) 0.325 
Cholesterol (mmol/L) 5.7 ± 1.6, 5.8 (4.6, 7.0) 6.6 ± 1.0, 6.6 (5.8, 7.4) 6.0 ± 1.3, 6.1 (5.3, 7.2) 5.5 ± 1.3, 5.1 (4.8, 6.4) 4.8 ± 1.8, 4.4 (3.7, 5.0) 6.0 ± 1.3, 5.9 (5.1, 6.9) 4.9 ± 0.0, 4.9 (4.9 , 4.9) <0.001 
HDL (mmol/L) 1.5 ± 0.4, 1.4 (1.2, 1.7) 1.6 ± 0.3, 1.6 (1.4, 1.8) 1.4 ± 0.5, 1.3 (1.0, 1.7) 1.4 ± 0.5, 1.2 (1.1, 1.7) 1.4 ± 0.4, 1.4 (1.2, 1.6) 1.5 ± 0.6, 1.3 (1.2, 1.6) 1.1 ± 0.0, 1.1 (1.1, 1.1) 0.112 
LDL (mmol/L) 4.0 ± 1.3, 4.0 (3.0, 5.0) 4.6 ± 0.8, 4.6 (4.0, 5.1) 4.4 ± 1.2, 4.4 (3.9, 5.0) 3.7 ± 0.8, 3.7 (3.1, 4.5) 3.2 ± 1.4, 2.7 (2.4, 3.2) 4.0 ± 1.3, 4.1 (3.1, 4.9) 3.6 ± 0.0, 3.6 (3.6, 3.6) <0.001 
Triglyceride (mmol/L) 1.9 ± 1.4, 1.5 (1.0, 2.2) 1.5 ± 0.5, 1.5 (1.1, 1.9) 1.8 ± 0.8, 1.7 (1.3, 1.9) 2.4 ± 2.1, 1.7 (1.1, 2.6) 1.7 ± 1.2, 1.2 (0.9, 2.2) 2.4 ± 1.7, 2.1 (1.1, 3.0) 1.9 ± 0.0, 1.9 (1.9, 1.9) 0.256 
Haemoglobin (mmol/L) 8.3 ± 1.6, 8.7 (7.9, 9.2) 8.9 ± 0.5, 8.9 (8.4, 9.2) 8.8 ± 1.5, 9.2 (8.7, 9.5) 8.2 ± 1.8, 8.6 (7.9, 9.2) 8.0 ± 2.1, 8.1 (7.0, 9.0) 7.7 ± 1.4, 8.2 (6.8, 8.9) 7.8 ± 0.0, 7.8 (7.8, 7.8) 0.006 
Leucocyte count (10e9/L) 8.0 ± 7.4, 6.9 (5.9, 8.2) 8.1 ± 11.1, 6.4 (5.4, 7.1) 10.0 ± 10.0, 6.6 (6.1, 9.5) 7.6 ± 2.6, 8.0 (6.8, 8.6) 7.5 ± 4.6, 6.9 (6.1, 7.8) 7.2 ± 1.3, 7.4 (6.2, 8.4) 9.2 ± 0.0, 9.2 (9.2, 9.2) 0.057 
Kidney biomarker level 
Uromodulin (ng/mL) 115.6 ± 101.0, 84.8 (37.0, 180.0) 260.4 ± 101.6, 227.6 (194.4, 277.8) 150.7 ± 54.8, 153.4 (121.5, 180.7) 116.7 ± 44.5, 111.8 (80.9, 139.4) 70.5 ± 47.1, 51.8 (36.1, 86.2) 29.1 ± 13.2, 30.3 (19.2, 38.0) 10.3 ± 7.0, 12.7 (6.7, 15.1) <0.001 
Cystatin C (nmol/L) 147.1 ± 93.8, 113.8 (72.7, 206.3) 66.7 ± 10.1, 68.2 (62.9, 73.4) 62.1 ± 10.9, 66.7 (55.4, 70.4) 91.5 ± 9.5, 89.1 (85.0, 98.9) 151.9 ± 27.4, 147.6 (127.0, 176.4) 273.8 ± 39.3, 271.1 (247.9, 304.3) 453.1 ± 70.5, 445.7 (408.2, 494.3) <0.001 
Creatinine (µmol/L) 176.9 ± 126.1, 127.0 (93.0, 223.3) 96.7 ± 22.1, 94.0 (80.0, 117.0) 82.4 ± 16.8, 76.9 (72.0, 93.0) 107.0 ± 21.8, 105.0 (91.5, 119.2) 169.3 ± 52.7, 160.9 (125.0, 206.0) 333.5 ± 131.5, 316.5 (246.1, 383.1) 669.5 ± 153.0, 670.0 (575.9, 763.3) <0.001 
BUN (mmol/L) 11.9 ± 8.0, 8.4 (5.9, 16.6) 5.9 ± 1.6, 5.7 (4.9, 6.8) 5.1 ± 1.2, 4.9 (4.6 , 5.9) 7.3 ± 3.4, 6.2 (6.0 , 7.4) 11.9 ± 4.4, 11.8 (7.9 , 15.5) 21.8 ± 6.5, 20.2 (16.7 , 26.7) 28.3 ± 4.3, 29.4 (26.0 , 31.2) <0.001 
eGFRHoek (mL/min) 64.2 ± 37.9, 57.7 (29.7, 92.9) 104.6 ± 20.7, 99.3 (91.9, 108.0) 114.0 ± 26.8, 101.0 (96.0, 123.2) 73.6 ± 7.8, 74.9 (67.1, 78.8) 43.7 ± 8.7, 43.6 (35.7, 51.1) 22.0 ± 3.8, 21.8 (19.1, 24.3) 11.8 ± 2.5, 12.0 (10.3, 13.4) <0.001 
eGFRCKD-EPI (mL/min) 49.4 ± 31.1, 44.0 (20.5, 71.5) 78.8 ± 17.5, 76.5 (65.1, 86.2) 92.8 ± 18.3, 88.9 (85.1, 97.0) 60.1 ± 11.5, 61.6 (51.5, 69.5) 32.5 ± 8.7, 31.4 (25.7, 39.5) 14.5 ± 4.0, 14.2 (11.6, 16.9) 6.7 ± 1.2, 7.0 (6.0, 7.5) <0.001 

NIDDM, non-insulin-dependent diabetes mellitus type II; IDDM, insulin-dependent diabetes mellitus type II; GN, glomerulonephritis: biopsy proven; pHPT, primary hyperparathyroidism; obstrN, obstructive nephropathy; toxN, CKD due to former nephrotoxic chemotherapy, HDL, high-density lipoprotein; LDL, low-density lipoprotein.

Table 3

Kidney allograft recipient’s demographics

Parameter Total (n = 44) Immediate graft function (n = 27) Delayed graft function (n = 17) P-value 
Age (years) 42.0 ± 10.8, 41.0 (33.8, 51.3) 43.1 ± 10.6, 44.0, (34.5, 52.0) 43.0 ± 11.0, 39.0, (32.0, 48.0) 0.421 
Gender     
 Male 27 (61.4) 16 (59.3) 11(64.7) 0.761 
 Female 17 (38.6) 11(40.7) 6 (35.3) 
Kind of donation     
 Deceased donor 44 (100) 27 (100) 17 (100) 1.000 
 Cold ischaemia time (min) 952.8 ± 455.3, 859.5 (609.8, 1227.0) 803.1 ± 347.9, 720.0 (570.5, 1005.0) 1190.6 ± 502.1, 1200.0 (794.0, 1500.0) 0.007 
 Maximum panel cytotoxicity (>5%) 11 (25.0) 7 (25.9) 4 (23.0) 1.000 
Immunosuppression—Induction therapy     
 ATG Fresenius, 9 mg/kg body weight intraoperatively (i.op.) 30 (68.2) 19 (70.4) 11 (64.5) 0.124 
 Lymphoglobulin Merieux, 30 mg/kg body weight i.op. 11 (25.0) 8 (29.6) 3 (17.6) 
 ATG Biotest, 1.5 mg/kg body weight i.op. 1 (2.3) 0 (0) 1 (5.9) 
 None 2 (4.5) 0 (0) 2 (11.8) 
Replacement (or additional therapies) therapy after Tx     
 None 27 (61.4) 27 (100) 0 (0) <0.001 
 Haemodialysis 16 (36.4) 0 (0) 16 (94.1) 
 Peritoneal dialysis 1 (2.3) 0 (0) 1 (5.9) 
 Number of dialysis sessions 9.1 ± 5.5 9.1 ± 5.5  
 Patient with rejection during the first year 16 (36.4) 8 (29.6) 8 (47.1) 0.337 
 Patient with Cytomegalovirus infections during the first year 8 (18.2) 4 (14.8) 4 (23.5) 0.690 
Biomarker levels at day of discharge     
 Day of discharge after Tx 40.7 ± 21.3, 37.5 (23.8, 50.3) 30.4 ± 13.2, 25.0 (20.0, 42.5) 56.9 ± 21.5, 51.0 (38.0, 63.0) 0.004 
 Uromodulin (ng/mL) 60.1 ± 43.1, 47.6 (29.3, 75.8) 76.6 ± 45.7, 63.0 (45.0, 95.1) 33.9 ± 19.1, 29.4 (21.2, 39.0) <0.001 
 Creatinine (µmol/L) 146.0 ± 53.5, 140.0, (107.8, 170.3) 124.7 ± 30.6, 118.0 (105.0, 145.5) 179.8 ± 63.7, 180.0, (161.0, 191.0) <0.001 
 eGFR (MDRD) (mL/min) 49.2 ± 17.0, 47.1 (36.7, 58.9) 55.4 ± 15.5, 53.6, (44.6, 66.5) 39.4 ± 14.5, 37.4 (32.3, 47.0) 0.004 
Parameter Total (n = 44) Immediate graft function (n = 27) Delayed graft function (n = 17) P-value 
Age (years) 42.0 ± 10.8, 41.0 (33.8, 51.3) 43.1 ± 10.6, 44.0, (34.5, 52.0) 43.0 ± 11.0, 39.0, (32.0, 48.0) 0.421 
Gender     
 Male 27 (61.4) 16 (59.3) 11(64.7) 0.761 
 Female 17 (38.6) 11(40.7) 6 (35.3) 
Kind of donation     
 Deceased donor 44 (100) 27 (100) 17 (100) 1.000 
 Cold ischaemia time (min) 952.8 ± 455.3, 859.5 (609.8, 1227.0) 803.1 ± 347.9, 720.0 (570.5, 1005.0) 1190.6 ± 502.1, 1200.0 (794.0, 1500.0) 0.007 
 Maximum panel cytotoxicity (>5%) 11 (25.0) 7 (25.9) 4 (23.0) 1.000 
Immunosuppression—Induction therapy     
 ATG Fresenius, 9 mg/kg body weight intraoperatively (i.op.) 30 (68.2) 19 (70.4) 11 (64.5) 0.124 
 Lymphoglobulin Merieux, 30 mg/kg body weight i.op. 11 (25.0) 8 (29.6) 3 (17.6) 
 ATG Biotest, 1.5 mg/kg body weight i.op. 1 (2.3) 0 (0) 1 (5.9) 
 None 2 (4.5) 0 (0) 2 (11.8) 
Replacement (or additional therapies) therapy after Tx     
 None 27 (61.4) 27 (100) 0 (0) <0.001 
 Haemodialysis 16 (36.4) 0 (0) 16 (94.1) 
 Peritoneal dialysis 1 (2.3) 0 (0) 1 (5.9) 
 Number of dialysis sessions 9.1 ± 5.5 9.1 ± 5.5  
 Patient with rejection during the first year 16 (36.4) 8 (29.6) 8 (47.1) 0.337 
 Patient with Cytomegalovirus infections during the first year 8 (18.2) 4 (14.8) 4 (23.5) 0.690 
Biomarker levels at day of discharge     
 Day of discharge after Tx 40.7 ± 21.3, 37.5 (23.8, 50.3) 30.4 ± 13.2, 25.0 (20.0, 42.5) 56.9 ± 21.5, 51.0 (38.0, 63.0) 0.004 
 Uromodulin (ng/mL) 60.1 ± 43.1, 47.6 (29.3, 75.8) 76.6 ± 45.7, 63.0 (45.0, 95.1) 33.9 ± 19.1, 29.4 (21.2, 39.0) <0.001 
 Creatinine (µmol/L) 146.0 ± 53.5, 140.0, (107.8, 170.3) 124.7 ± 30.6, 118.0 (105.0, 145.5) 179.8 ± 63.7, 180.0, (161.0, 191.0) <0.001 
 eGFR (MDRD) (mL/min) 49.2 ± 17.0, 47.1 (36.7, 58.9) 55.4 ± 15.5, 53.6, (44.6, 66.5) 39.4 ± 14.5, 37.4 (32.3, 47.0) 0.004 

Data are presented as mean ± SD and median with IQR in brackets or as absolute numbers (n) with percentage (%) in brackets. Tx, transplantation; MDRD, Modification of Diet in Renal Disease.

Definition of CKD stages and outcome parameters

Based on clinical and eGFR data patients were assigned to one of five stages of CKD. The underlying kidney diseases were histologically characterized by routine stains [haematoxylin & eosin (HE), periodic acid-Schiff (PAS)], immunohistology and ultrastructure of renal biopsy sections. GFR was determined by eGFRCysC. Patients were assigned to CKD stages according to Kidney Disease: Improving Global Outcomes (KDIGO) guidelines: CKD 1 with eGFR >90 mL/min, CKD 2 61–90 mL/min, CKD 3 31–60 mL/min, CKD 4 16–30 mL/min and CKD 5 0–15 mL/min. We included 33 patients without kidney disease as the control group (non-CKD).

Forty-four patients had undergone KTX between 1990 and 1995 at the Berlin-Friedrichshain Kidney Centre, Germany. Follow-up was until 2002, unless death occurred. Clinical characteristics and definitions of IGF and DGF are summarized in Table 3.

Antibodies

Mouse moAb were generated against purified human uUmod. BALB/c mice were immunized with emulsified purified Umod in ABM-S adjuvant. After booster injections (ABM-N adjuvant) mice were anesthetized, splenectomized and spleen cells were PEG fused with mouse X63Ag8.653 myeloma cells. Hybridoma supernatants were screened for specific antibodies by ELISA with purified Umod as antigen. Positive hybridomas were subcloned by limiting dilution. Distinct hybridomas were expanded and purified. Specificity of moAb was assessed by immunocytochemistry using paraffin-embedded tissue sections from human kidney, liver, intestine, muscle and lymph nodes. Only human kidney showed a positive histochemical reaction with anti-Umod antibodies.

Uromodulin ELISA

Purified IgG1 of moAb T112A12 for coating the microtiter plates and moAb T5.G as detection antibody were used to establish an ELISA. All analytical data (sensitivity, precision, linearity, etc.) for CE labelling were determined and are available from the manufacturer (Euroimmun). All measurements were done with a 1:101 serum dilution in sample buffer. The linear measurement range was 2–400 ng/mL.

Immunohistological and cytological analyses

Normal renal tissue was obtained from kidney resection in renal carcinoma patients and fixed in paraformaldehyde. The tissue was cut into small blocks and processed in an automatic freeze substitution unit (Reichert-Jung AFS, Leica Microsystems, Wetzlar, Germany). Umod moAb clonotypes 109F9 (IgG1), 112A12 (IgG1) and 5A2 (IgG2b) were used for staining.

Immunocytochemistry

After primary antibody incubation deparaffinized sections were processed with secondary biotinylated antibody and avidin-peroxidase complex according to standard procedures. Biopsy material was obtained from patients (21 men, 11 women, mean age 54.7 ± 17 years) with kidney diseases including IgA nephropathy (n = 12), membranous glomerulonephritis (mGN) (n = 8), ANCA-positive RPGN (n = 12), where a minimum of two to three corticomedullary nephrons were noticed, displaying interstitial fibrosis and tubular atrophy of  more than 10%. In addition, we assessed sections of non-functioning allografts of nine patients (one female, eight males). Staining included HE, Masson-Trichrom, Giemsa and PAS. Sections were examined for Umod location by two different pathologists in a blinded manner.

Immunoelectron microscopy

Sections were incubated with the primary antibody, followed by gold-conjugated second antibody. Stained sections were studied in a Philips/FEI CM100 transmission electron microscope. In controls, sections were incubated with secondary antibodies alone or with non-specific moAb. None of the controls showed any labelling.

Statistical analysis

All statistical analyses were performed with SigmaPlot 13 and the language R for statistical computing. Continuous data are expressed as mean with standard deviation (SD) and median with interquartile range (IQR). Categorical variables are reported in absolute numbers and percentages. Differences concerning continuous parameters and biomarker concentrations between two groups were analysed using non-parametric Mann–Whitney U-test. The Kruskal–Wallis test was used for multiple group comparison. For categorical variables, possible differences between groups were examined using non-parametric chi-square test (Fisher's exact test). The association between sUmod and creatinine, cystatin C and eGFR was determined by Spearman’s rank correlation analysis. Multiple linear regression modelling was conducted to evaluate the association of Umod (independent variable) and eGFRCysC (dependent variable) adjusted for gender and age. As part of the regression modelling sUmod and eGFR were logarithmically transformed. Wilcoxon-rank test and/or receiver-operating characteristic (ROC) analysis were used to evaluate diagnostic validity of sUmod, cystatin C, creatinine, BUN and eGFRCysC to differentiate between non-CKD and CKD stage 1. Relative risk (RR) and odds ratio (OR) for DGF after KTX were calculated on the basis of mean cut-off (day 7–25 post-transplant). Daily cut offs were determined according to Youden’s index. The probability of error (type 1) of a two tailed P-value <0.05 was considered to be statistically significant.

RESULTS

sUmod stability

sUmod stability was analysed over 4 weeks (

). Serum samples from eight individuals were stored at either −20°C, 4°C or 37°C before monitoring Umod levels. A high concordance in Umod concentrations was observed for all three storage conditions with mean deviations of 4% (−20 versus 4°C) and 10% (−20 versus 37°C). Even up to five freeze/thaw cycles had no impact on sUmod recoveries ().

sUmod expected normal values

sUmod was assessed in 190 blood donors (89 males, 101 females, age 18–60 years, median age 40 years). Donors were checked to be healthy by clinical examination and laboratory tests (normal controls, Table 1). Females revealed higher sUmod than males [females: median 230 ng/mL (IQR: 165, 308); males: median 188 ng/mL (IQR: 44, 251), P = 0.006]. In adults, no correlation between age and sUmod was observed (rs = −0.0197, P = 0.787). sUmod in children and adolescents (n = 443, age 1–17 years, median age 6 years) was comparable to values found in adults [median 193 ng/mL (IQR: 143, 254)]. sUmod levels did not differ between boys and girls (P = 0.389).

sUmod in renal failure

sUmod levels and the filtration markers creatinine and cystatin C were analysed in 165 patients (110 males, 55 females, age 22–94 years, median age 65 years) with various diseases potentially affecting the kidney. sUmod ranged from 1 to 667 ng/mL with a median of 85 ng/mL (Table 2). In univariate analysis, sUmod was significantly correlated with cystatin C, creatinine and all eGFR calculations. Scatter plots demonstrate an inverse hyperbolic relationship of sUmod with cystatin C (rs = −0.862, P < 0.001), creatinine (rs = −0.802, P < 0.001) and blood urea nitrogen (BUN, rs = −0.645, P < 0.001) and, accordingly, a positive relationship with cystatin C-based eGFR (eGFRCysC, Hoek’s equation) (rs  = 0.862, P < 0.001) and eGFR (CKD-EPI) (rs  = 0.842, P < 0.001) (Figure 1). The creatinine–sUmod ratio was not superior compared with assessing sUmod levels alone. The data disclose that sUmod parallels the progressive deterioration of kidney function and behaves oppositely to the filtration markers. In a multiple linear regression model, sUmod was adjusted for age and gender, which did not significantly influence the regression model with eGFRCysC (log) as a dependent variable. The correlation between sUmod (log) and eGFRCysC (log) was r2 =0.735 (P < 0.001).

FIGURE 1

Relationship between serum uromodulin and creatinine (A), blood urea nitrogen (B), cystatin C (C) and eGFRcystC (D). rs: Spearman’s rank correlation coefficient.

FIGURE 1

Relationship between serum uromodulin and creatinine (A), blood urea nitrogen (B), cystatin C (C) and eGFRcystC (D). rs: Spearman’s rank correlation coefficient.

To study the relationship between sUmod and GFR, sUmod levels were assigned to the stages of renal failure CKD 1–5, based on the eGFRCysC data (Figure 2). sUmod decreased with progressive CKD (non-CKD: median 228 ng/mL; stage 1: 153 ng/mL, stage 2: 107 ng/mL; stage 3: 52 ng/mL; stage 4: 30 ng/mL; stage 5: 13 ng/mL). Differences in sUmod were shown for all pairs of CKD stages (P < 0.001), except for stages 1 and 2 (P = 0.051). The most prominent changes, i.e. the steepest decline in sUmod, occurred within the ‘creatinine blind range’ (Figure 2).

FIGURE 2

Serum uromodulin concentration of patients without kidney diseases (non-CKD) and patients with CKD (stage 1 to stage 5) based on eGFRcysC assignment.

FIGURE 2

Serum uromodulin concentration of patients without kidney diseases (non-CKD) and patients with CKD (stage 1 to stage 5) based on eGFRcysC assignment.

All traditional kidney markers significantly differentiated between all CKD stages from 1 to 5, but none showed a conclusive statistical difference between non-CKD and CKD 1. Creatinine and cystatin C were even higher in non-CKD compared with CKD stage 1 [creatinine: median 94.0 µmol/L (IQR: 80.0, 117.0) versus 76.9 µmol/L (72.0, 93.0), P = 0.011; cystatin C: median 68.2 nmol/L (IQR: 62.9, 73.4) versus 66.7 nmol/L (IQR: 55.4, 70.4), P =  0.153].

ROC analysis of sUmod to differentiate between non-CKD and CKD 1 resulted in an AUC of 0.90 (P < 0.001, 95% CI 0.808 vs 0.990) at an optimal cut-off of 163 ng/mL (sensitivity 57.1%, specificity 100%). None of the other parameters including eGFR discriminated between stages non-CKD and CKD 1 (cystatin C: AUC 0.39; creatinine: AUC 0.29; BUN: AUC 0.37; eGFRCysC: AUC 0.39; eGFR (CKD-EPI): AUC 0.26) (Figure 3).

FIGURE 3

ROC analysing the ability of different parameters to differentiate patients without kidney disease (non-CKD) and CKD stage 1; BUN =  blood urea nitrogen, eGFR = estimated glomerular filtration rate, AUC =  area under the curve.

FIGURE 3

ROC analysing the ability of different parameters to differentiate patients without kidney disease (non-CKD) and CKD stage 1; BUN =  blood urea nitrogen, eGFR = estimated glomerular filtration rate, AUC =  area under the curve.

sUmod in autoimmune diseases

Three patient cohorts with autoimmune diseases, including pulmorenal syndromes (n = 10), SLE (n = 53) and PLA2-R Ab-positive glomerulonephritis (mGN, n = 20) were assessed. Patients with acute anti-GBM nephritis of CKD 3–5 had much lower sUmod compared with blood donors (median 37 ng/mL, P < 0.001,

, Figure S1). Patients with SLE-nephritis (CKD 2–4) had median sUmod of 101 ng/mL, P < 0.001, while those without kidney involvement had similar sUmod levels compared with blood donors (median: 259 ng/mL, P = 0.485). Patients with mGN had median sUmod of 81 ng/mL, P < 0.001.

sUmod after KTX

As shown in Figure 4 prior to surgery sUmod was not detectable in all patients due to end-stage renal disease (ESRD). Immediately after transplantation sUmod rapidly increased in recipients with IGF and DGF until post-KTX day 5, reaching 43 ng/mL (median) in IGF and 33 ng/mL (median) in DGF, respectively. Thereafter, levels diverged from day 7 (IGF: 59.5 ± 41.2 ng/mL, DGF: 33.1 ± 20 ng/mL; P = 0.033) to day 25. The cohort with DGF failed to reach sUmod levels above 40 ng/mL (cut off 39.26 ng/mL). From day 5 on, the RR and OR to develop DGF rose dramatically on the basis of sUmod (day 5: RR 1.3, OR 1.5; day 9: RR 2.9, OR 6.4, etc.; day 25: RR 8.2, OR 26;

, Table S2). Sixteen of 17 patients with DGF needed 3–24 postoperative haemodialysis treatments. Patients with IGF were discharged from hospital after a median of 25 days, with a median creatinine of 118 µmol/L and a stable ‘high’ sUmod (76.6 ± 46.6 ng/mL). Patients with DGF and additional complications were discharged from hospital much later, at a median of 51 days, with a median creatinine of 180 µmol/L and lower sUmod concentration (33.9 ± 19.7 ng/mL, P < 0.001).
FIGURE 4

Post-kidney transplant comparison of serum uromodulin and creatinine in patients with immediate and delayed graft function.

FIGURE 4

Post-kidney transplant comparison of serum uromodulin and creatinine in patients with immediate and delayed graft function.

Umod distribution in normal human kidney

Labelling with anti-Umod antibodies was confined to the distal tubule (Figure 5). Umod was seen throughout the cytoplasm and in apical and basolateral membranes of the distal tubule cells. The images illustrate the transition from the unlabelled thin ascending limb to the intensely labelled TAL. In the macula densa, the cells were largely unlabelled or weakly labelled, whereas the other distal tubule cells were intensely labelled.

FIGURE 5

Immunohistochemical distribution of Umod in the normal human kidney (ac); semithin cryosections. (a) Apical, basolateral and cytoplasmatic staining, (b) transition of the thin ascending limb (negative) to the TAL (positive, arrows), (c) macula densa reveals a faint staining pattern only (arrowheads).

FIGURE 5

Immunohistochemical distribution of Umod in the normal human kidney (ac); semithin cryosections. (a) Apical, basolateral and cytoplasmatic staining, (b) transition of the thin ascending limb (negative) to the TAL (positive, arrows), (c) macula densa reveals a faint staining pattern only (arrowheads).

Umod location in kidney diseases

Umod staining of biopsy material from patients with proven kidney diseases and of grafts that had to be explanted due to ESRD showed an irregular pattern and increased intracellular expression of the antigen compared with healthy controls (

). In kidney biopsies, TAL segments were under-represented compared with other tubular segments and were not accessible to further morphometric analyses. In most Umod-positive cells, the antigen appeared to accumulate and condense around the nuclei, the endoplasmatic reticulum and the Golgi apparatus. The cytoplasma of TAL cells from explanted kidney grafts stained for Umod with a dramatically increased intensity; here potential changes in Umod expression of apical and basolateral membranes could not be assessed ().

Immunoelectron microscopy

Immunogold labelling of Umod was seen at the apical and basolateral membrane, in the cytoplasma, the granular endoplasmic reticulum, the Golgi apparatus and in vesicles located apically but also in vesicles close to the basement membrane (Figure 6). Labelling was largely confined to the external layer of the apical and external parts of the basolateral plasma membrane including their infoldings. No labelling was seen in the collecting duct either in principal or in intercalated cells.

FIGURE 6

Immunogold labelling of Umod revealing the transport of the antigen within cytoplasmatic vesicles (small arrowheads) to the apical (a) as well as to the basolateral cell pole (b). Shuttle of uromodulin (c) from the endoplasmatic reticulum (RER) to the Golgi apparatus (G) (inset to C), through cytoplasmic vesicles (arrowheads). Basolateral membrane (BM) labelling (arrow). M, mitochondria.

FIGURE 6

Immunogold labelling of Umod revealing the transport of the antigen within cytoplasmatic vesicles (small arrowheads) to the apical (a) as well as to the basolateral cell pole (b). Shuttle of uromodulin (c) from the endoplasmatic reticulum (RER) to the Golgi apparatus (G) (inset to C), through cytoplasmic vesicles (arrowheads). Basolateral membrane (BM) labelling (arrow). M, mitochondria.

DISCUSSION

The majority of studies on Umod to date focus on the urinary form of Umod. sUmod in contrast has rarely been investigated [25, 26] and its clinical significance, molecular structure and function in blood are poorly understood. Brezis et al. first emphasized that the renal medulla is a neglected tissue compartment, not adequately reached by renal biopsy. Umod is a major secreted protein produced by the renal medulla and experimental data support the pivotal role of medullary hypoxic injury in kidney diseases [12, 18–23, 27].

We assessed sUmod in patients with different CKD stages by ELISA. sUmod levels inversely correlated with creatinine and creatinine/cystatin C-based eGFR. The lowest sUmod was measured in patients with CKD stages 4 and 5. We speculate that low sUmod results from a disturbed biosynthesis in cells of the TAL. Patients with renal injury also show a decrease in urinary Umod levels [28, 32–34]. sUmod is not measurable in anephric patients [24]. In elderly patients, sUmod displayed inverse relationships with creatinine, cystatin C and urea, and, correspondingly, a positive relation with eGFR [26]. Contradicting all recent data, Prajczer et al. report that sUmod is very low in healthy controls and increased in patients with CKD 1–5 [25]. We can only speculate that these conflicting results depend on the design of their ELISA and potential additional preanalytic factors.

Addressing the limitations of our study we did not measure GFR by inulin or iohexol clearance. Instead we calculated eGFR on a creatinine and cystatin C basis, which is presently still the most accepted method under clinical routine conditions.

With the help of our sensitive assay, to the best of our knowledge, we first demonstrated Umod as a physiological and very stable antigen circulating in the blood of healthy adults and children without any known kidney problems.

Importantly, we defined conspicuous sUmod at values below the fifth percentile range of healthy individuals and found different thresholds for children and males (<80 ng/mL) and for females (<100 ng/mL). Higher sUmod levels in healthy female compared with men were also observed in a previous study [26] and parallel similar sex differences in the urinary excretion of Umod and other tubular markers [43–46]. Although gender-dependent endocrine, haemodynamic, metabolic and genetic factors, expression of receptors, as well as sex differences in the susceptibility of cell integrity may contribute to this observation, most questions are still unanswered [47, 48].

The clinical data presented here are limited to patients with CKD (apart from the cohorts after grafting) and need to be extended to cases suffering from AKI, which are now under study. In contrast to Risch et al. [26] we could not find a correlation between sUmod and age (age range from 18 to 60 years; r = −0.60, P < 0.001). A possible explanation is that Risch et al. focused on elderly participants (mean age 71 ± 7 years) and the physiological decrease of sUmod probably starts at the age of 60 years. Age-related changes in the healthy kidney assessed by multidetector CT scans showed diverged alterations (disequilibrium) of renal compartments, where cortical volume (and GFR in parallel) declines, whereas medullary parenchymal volume increases [49]. This suggests that synthesis and release of Umod by TAL into the blood might be maintained over a longer period.

According to as yet unpublished studies, we solely found monomeric sUmod and so far did not observe any aggregation processes in serum. As a key novel finding we noticed extraordinarily stable signals over weeks even at increased temperatures. In contrast, Umod from urine (uUmod) lacks stability depending on the storage conditions, centrifugation, vortexing, pH, electrolytes and osmolality [31, 43]. The changing conformation of uUmod and the varying degree of polymerization, possibly modulating antigenic binding sites, may impede the reproducibility of the results.

A major finding is the comparative analysis of sUmod with creatinine, cystatin C and eGFR indicating that lower sUmod reflects a decline in kidney function and may precede the onset of CKD, as was shown by us in another patient cohort [50].

Tubulointerstitial injury is closely linked to progression of renal disease [51, 52]. We assume that low sUmod parallels a structural derangement of the TAL. sUmod levels may drop due to altered intracellular processing and reduced basolateral exocytosis and mirror the reduced quantity of intact TAL cells. Acute and chronic renal injury characterized by epithelial atrophy and tubular loss of both cortical and medullary segments may also parallel changes of the vascular microarchitecture [53]. Interstitial vascular remodelling may influence the structural integrity of tubular cells, e.g. vulnerable TAL segments, and change glomerular morphology and haemodynamics, which would link sUmod levels with eGFR (CKD stages). In the present study, we could not assess the relation between histological alterations of Umod-expressing cells and sUmod. The amount of TAL segments varied greatly among the biopsies, and Umod-positive cells appear considerably under-represented in renal clinical pathology.

One known secretory pathway of Umod occurs at the apical membrane of the TAL cells. Umod concentrations in urine are approximately 1000-fold higher than in serum. Umod’s apical secretion is based on exocytosis and the formation of a GPI anchor, followed by its release from the membrane via proteolytic cleavage [3, 27]. However, besides the luminal secretion process of Umod an additional basolateral pathway was suggested [54, 55]. Our data confirm the view that Umod is synthesized in the endoplasmatic reticulum, transferred to the Golgi apparatus and transported to the apical membrane by vesicles. Finally it sheds into the luminal space. In addition, we clearly demonstrated a second, basolateral pathway of Umod secretion (Figure 6), which apparently contributes to sUmod levels (Figure 7).

FIGURE 7

Cartoon demonstrating the proposed bidirectional physiological transfer of Umod to the urinary space and to the tubulointerstitial space. As shown in Figure 6b, Umod is physiologically released from basolateral plasma membranes and membrane infoldings into the interstitium and enters the blood compartment (a). Umod serum levels are correlated with eGFR. sUmod concentrations may depend on the cellular integrity of the TAL segment. Therefore, sUmod might be released into serum at a decreased rate in renal injury (b).

FIGURE 7

Cartoon demonstrating the proposed bidirectional physiological transfer of Umod to the urinary space and to the tubulointerstitial space. As shown in Figure 6b, Umod is physiologically released from basolateral plasma membranes and membrane infoldings into the interstitium and enters the blood compartment (a). Umod serum levels are correlated with eGFR. sUmod concentrations may depend on the cellular integrity of the TAL segment. Therefore, sUmod might be released into serum at a decreased rate in renal injury (b).

We interpret sUmod as an early indicator defining the integrity and the functional viability of the TAL. In contrast, the traditional endogenous filtration markers are not related to defined renal structures and need mathematical transformation. Established renal markers have limitations in detecting early or slight changes in kidney function, particularly in a range when approximately half of the filtration capacity is lost [56]. Following the analyses and conclusions from experimental animal data on TAL segments [18–23] we assume that sUmod levels may somehow reflect the amount of intact nephrons being linked to the ‘functioning kidney mass’.

Another important new finding is that, in accordance with eGFR, sUmod recovers to higher levels as shown in transplant patients, and may distinguish IGF from DGF. Worsening of graft function was always accompanied by long-lasting reduced sUmod levels and predictable after the seventh postoperative day. It is a subject of future research to clarify whether low sUmod levels in patients with DGF are predictive for later chronic graft dysfunction and serious patient outcome [57].

In conclusion, we have demonstrated that sUmod concentrations decrease in line with eGFR and with biopsy-classified stages of renal failure. Our data strengthen the link between sUmod and kidney disease and suggest that sUmod might be superior compared with the traditional filtration markers to detect early stages of CKD, independent of the underlying disease. sUmod should not only be interpreted as a complementary parameter describing global ‘kidney function’, but may shift our focus to potential structural and metabolic lesions of TAL cells. These encourage more extensive elucidation of the role of sUmod as a diagnostic tool in comparison with, and likely in concert with, other novel renal biomarkers [58–61].

SUPPLEMENTARY DATA

are available online at http://ndt.oxfordjournals.org.

ACKNOWLEDGEMENTS

We thank Prof. Walter Nathrath, Institute of Pathology, Prof. Oliver Reich, Department of Urology, Klinikum Harlaching, LMU München, and Dr Florian Laenger, Institute of Pathology, Goethe-University Frankfurt am Main and Hannover Medical School, for supplying paraffin-embedded material and analysing tissue staining patterns. Parts of the work were presented in abstract form at the congress of the German Society of Nephrology, Berlin, September 2015, and the congress of Deutsche Transplantationsgesellschaft, Dresden, October 2015, and the 10th Intenational Congress on Autoimmunity, Leipzig, April 2016. We thank Prof. Seymour Rosen, Beth Israel Deaconess MC, Harvard Medical School, Boston, Prof. Anna Koettgen, Medizinische Klinik IV, Department of Nephrology, University of Freiburg, and Dr Ulrike Krummrei, Euroimmun, for critical proofreading of the manuscript, and Prof. Olivier Devuyst, University Zürich, Switzerland, and Prof. Hans-Joachim Anders, LMU Munich for collaboration and valuable discussions.

CONFLICT OF INTEREST STATEMENT

J.E.S. received travel support from Hoffmann-La Roche and Euroimmun, R.G. received travel support from Roche and ThermoScientific, V.H., M.B. and W.S. are full time employees of Euroimmun and J.K. is member of the supervisory board of Euroimmun. W.A.N., E.I.C. and H.S. declare no potential conflicts of interest.

REFERENCES

1
Tamm
I
Horsfall
FL.
Characterization and separation of an inhibitor of viral hemagglutination present in urine
.
Proc Soc Exp Biol Med
 
1990
;
74
:
106
108
2
Devuyst
O
Dahan
K
Pirson
Y.
Tamm–Horsfall protein or uromodulin: new ideas about an old molecule
.
Nephrol Dial Transplant
 
2005
;
20
:
1290
1294
3
Rampoldi
L
Scolari
F
Amoroso
A
et al.  .
The rediscovery of uromodulin (Tamm–Horsfall protein): from tubulointerstitial nephropathy to chronic kidney disease
.
Kidney Int
 
2011
;
80
:
338
347
4
Mount
DB.
Thick ascending limb of the loop of Henle
.
Clin J Am Soc Nephrol
 
2014
;
9
:
1974
1986
5
Cavallone
D
Malagolini
N
Serafini-Cessi
F.
Mechanism of release of urinary Tamm–Horsfall glycoprotein from the kidney GPI-anchored counterpart
.
Biochem Biophys Res Commun
 
2001
;
280
:
110
114
6
Pak
J
Pu
X
Zhang
ZT
et al.  .
Tamm–Horsfall protein binds to type 1 fimbriated Escherichia coli and prevent E. coli from binding to uroplakin Ia and Ib receptors
.
J Biol Chem
 
2001
;
276
:
9924
9930
7
Raffi
HS
Bates
JM
Jr
Laszik
Z
et al.  .
Tamm–Horsfall protein protects against urinary tract infection by Proteus mirabilis
.
J Urol
 
2009
;
181
:
2332
2338
8
Serafini-Cessa
F
Monti
A
Cavallone
D.
N-glycans carried by Tamm–Horsfall glycoprotein have a crucial role in the defense against urinary tract diseases
.
Glycoconj J
 
2005
;
22
:
383
394
9
Mo
L
Huang
HY
Zhu
XH
et al.  .
Tamm–Horsfall protein is a critical renal defense factor protecting against calcium oxalate crystal formation
.
Kidney Int
 
2004
;
66
:
1159
1166
10
Graham
LA
Padmanabhan
S
Fraser
NJ
et al.  .
Validation of Uromodulin as a candidate gene for human essential hypertension
.
Hypertension
 
2014
;
63
:
551
558
11
Zacchia
M
Capasso
G.
The importance of uromodulin as regulator of salt reabsorption along the thick ascending limb
.
Nephrol Dial Transplant
 
2015
;
30
:
158
160
12
Trudu
M
Janas
S
Lanzani
C
et al.  .
Common noncoding UMOD gene variants induce salt-sensitive hypertension and kidney damage by increasing uromodulin expression
.
Nat Med
 
2013
;
19
:
1655
1660
13
Scolari
F
Izzi
C
Ghiggeri
GM.
Uromodulin: from monogenic to multifactorial diseases
.
Nephrol Dial Transplant
 
2015
;
30
:
1250
1256
14
Kreft
B
Jabs
WJ
Laskay
T
et al.  .
Polarized expression of Tamm–Horsfall protein by renal tubular epithelial cells activates human granulocytes
.
Infect Immun
 
2002
;
70
:
2650
2656
15
Iorember
FM
Vehaskari
VM.
Uromodulin: old friend with new roles in health and disease
.
Pediatr Nephrol
 
2014
;
29
:
1151
1158
16
Säemann
MD
Weichhart
T
Zeyda
M
et al.  .
Tamm–Horsfall glycoprotein links innate immune cell activation with adaptive immunity via a Toll-like receptor-4-dependent mechanism
.
J Clin Invest
 
2005
;
115
:
468
475
17
Darisipudi
MN
Thomasova
D
Mulay
SR
et al.  .
Uromodulin triggers IL-1ß-dependent innate immunity via the NLRP3 inflammasome
.
J Am Soc Nephrol
 
2012
;
23
:
1783
1789
18
Schley
G
Klanke
B
Schödel
J
et al.  .
Hypoxia-inducible transcription factors stabilization in the thick ascending limb protects against ischemic acute kidney injury
.
J Am Soc Nephrol
 
2011
;
22
:
2004
2015
19
Brezis
M
Rosen
S.
Hypoxia of the renal medulla—its implications for disease
.
N Engl J Med
 
1995
;
332
:
647
655
20
Heyman
SN
Rosenberger
C
Rosen
S.
Experimental ischemia—reperfusion: biases and myths—the proximal vs. distal hypoxic tubular injury debate revisited
.
Kidney Int
 
2010
;
77
:
9
16
21
Rosen
S
Stillman
IE.
Acute tubular necrosis is a syndrome of physiologic and pathologic dissociation
.
J Am Soc Nephrol
 
2008
;
19
:
871
875
22
Srichai
MB
Hao
C
Davis
L
et al.  .
Apoptosis of the thick ascending limb results in acute kidney injury
.
J Am Soc Nephrol
 
2008
;
19
:
1538
1546
23
Sharfuddin
AA
Molitoris
BA.
Pathophysiology of ischemic acute kidney injury
.
Nat Rev Nephrol
 
2011
;
7
:
189
200
24
Thornley
C
Dawnay
A
Cattell
WR.
Human Tamm–Horsfall glycoprotein: urinary and plasma levels in normal subjects and patients with renal disease determined by a fully validated radioimmunoassay
.
Clin Sci (Lond)
 
1985
;
68
:
529
535
25
Prajczer
S
Heidenreich
U
Pfaller
W
et al.  .
Evidence for a role of uromodulin in chronic kidney disease progression
.
Nephrol Dial Transplant
 
2010
;
25
:
1896
1903
26
Risch
L
Lhotta
K
Meier
D
et al.  .
The serum uromodulin level is associated with kidney function
.
Clin Chem Lab Med
 
2014
;
52
:
1755
1761
27
El-Achkar
TM
McCracken
R
Liu
Y
et al.  .
Tamm–Horsfall protein translocates to the basolateral domain of thick ascending limbs, interstitium, and circulation during recovery from acute kidney injury
.
Am J Physiol Renal Physiol
 
2013
;
304
:
F1066
F1075
28
Kaden
J
Groth
J
May
G
et al.  .
Urinary Tamm–Horsfall protein as a marker of renal transplant function
.
Urol Res
 
1994
;
22
:
131
136
29
Kobayashi
K
Fukuoka
S.
Conditions for solubilization of Tamm–Horsfall protein/uromodulin in human urine and establishment of a sensitive and accurate enzyme-linked immunosorbent assay (ELISA) method
.
Arch Biochem Biophys
 
2001
;
388
:
113
120
30
Catalano
C
Torffvit
O.
Urinary excretion of Tamm–Horsfall protein in normotensive, normo-albuminuric type 1 diabetic patients
.
Nephron
 
1996
;
72
:
436
441
31
Youhanna
S
Weber
J
Beaujean
V
et al.  .
Determination of uromodulin in human urine: influence of storage and processing
.
Nephrol Dial Transplant
 
2014
;
29
:
136
145
32
Shlipak
MG
Li
Y
Fox
C
et al.  .
Uromodulin concentrations are not associated with incident CKD among persons with coronary artery disease
.
BMC Nephrol
 
2011
;
12
:
2
33
Zhou
J
Chen
Y
Liu
Y
et al.  .
Urinary uromodulin excretion predicts progression of chronic kidney disease resulting from IgA nephropathy
.
PLoS One
 
2013
;
8
:
e71023
34
Garimella
PS
Biggs
ML
Katz
R
et al.  .
Urinary uromodulin, kidney function, and cardiovascular disease in elderly adults
.
Kidney Int
 
2015
;
88
:
1126
1134
35
Köttgen
A
Hwang
SJ
Larson
MG
et al.  .
Uromodulin levels associate with common UMOD variant and risk for incident CKD
.
J Am Soc Nephrol
 
2010
;
21
:
337
344
36
Olden
M
Corre
T
Hayward
C
et al.  .
Common variants in UMOD associate with urinary uromodulin levels: a meta-analysis
.
J Am Soc Nephrol
 
2014
;
25
:
1869
1882
37
Kumar
S.
Mechanisms of injury in Uromodulin-associated kidney disease
.
J Am Soc Nephrol
 
2007
;
18
:
10
12
38
Bleyer
AJ
Hart
PS
Kmoch
S.
Hereditary interstitial kidney disease
.
Semin Nephrol
 
2010
;
30
:
366
373
39
Ekici
AB
Hackenbeck
T
Morinière
V
et al.  .
Renal fibrosis is the common feature of autosomal dominant tubulointerstitial kidney diseases caused by mutations in mucin 1 or uromodulin
.
Kidney Int
 
2014
;
86
:
589
599
40
Dahan
K
Devuyst
O
Smaers
M
et al.  .
A cluster of mutations in the UMOD gene causes familial juvenile hyperuricemic nephropathy with abnormal expression of uromodulin.
J Am Soc Nephrol
 
2003
;
14
:
2883
2893
41
Hart
TC
Gorry
MC
Hart
PS
et al.  .
Mutations of the UMOD gene are responsible for medullary cystic kidney disease 2 and familial juvenile hyperuricaemic nephropathy
.
J Med Genet
 
2002
;
39
:
882
892
42
Reznichenko
A
Böger
CA
Snieder
H
et al.  .
UMOD as a susceptibility gene for end-stage renal disease
.
BMC Med Genet
 
2012
;
13
:
78
43
Pruijm
M
Ponte
B
Ackermann
D
et al.  .
Association of urinary Uromodulin with clinical characteristics and markers of tubular function in the general population
.
Clin J Am Soc Nephrol
 
2016
;
11
:
70
80
44
Kaden
J.
Optimal management of induction therapy with ATG in kidney allograft rejection
.
Int J Immunother
 
1999
;
15
:
115
124
45
Glauser
A
Hochreiter
W
Jaeger
P
et al.  .
Determinants of urinary excretion of Tamm–Horsfall protein in non-selected kidney stone formers and healthy subjects
.
Nephrol Dial Transplant
 
2000
;
15
:
1580
1587
46
Trachtenberg
F
Barregard
L.
The effect of age, sex, and race on urinary markers of kidney damage in children
.
Am J Kidney Dis
 
2007
;
50
:
938
945
47
Neugarten
J
Golestaneh
L.
Gender and the prevalence and progression of renal disease
.
Adv Chronic Kidney Dis
 
2013
; 20:
390
395
48
Silbiger
SR.
Raging hormones: gender and renal disease
.
Kidney Int
 
2011
;
79
:
382
384
49
Wang
X
Vrtiska
TJ
Avula
RT
et al.  .
Age, kidney function, and risk factors associate differentially with cortical and medullary volumes of the kidney
.
Kidney Int
 
2014
;
85
:
677
685
50
Steubl
D
Block
M
Herbst
V
et al.  .
Plasma uromodulin correlates with kidney function and identifies early stages in chronic kidney disease patients
.
Medicine
 
2016
;
95
:
e3011
51
Nath
KA.
Tubulointerstitial changes as a major determinant in the progression of renal damage
.
Am J Kidney Dis
 
1992
;
20
:
1
17
52
Healy
E
Brady
HR.
Role of tubule epithelial cells in the pathogenesis of tubulointerstitial fibrosis induced by glomerular disease
.
Curr Opin Nephrol Hypertens
 
1998
;
7
:
525
530
53
Farris
AB
Ellis
CL
Rogers
TE
et al.  .
Renal medullary and cortical correlates in fibrosis, epithelial mass, microvascularity, and microanatomy using whole slide image analysis morphometry
.
PLoS One
 
11
:
e0161019
54
El-Achkar
TM
Wu
XR.
Uromodulin in kidney injury: an instigator, bystander, or protector?
Am J Kidney
 
2012
;
59
:
452
461
55
Jennings
P
Aydin
S
Kotanko
P
et al.  .
Membrane targeting and secretion of mutant uromodulin in familial juvenile hyperuricemic nephropathy
.
J Am Soc Nephrol
 
2007
;
18
:
264
273
56
Biesen van
W
Nagler
EV.
A Swiss army knife for estimating kidney function: why new equations will not solve the real problem
.
Nephrol Dial Transplant
 
2016
;
31
:
685
687
57
Gill
J
Dong
J
Rose
C
et al.  .
The risk of allograft failure and the survival benefit of kidney transplantation are complicated by delayed graft function
.
Kidney Int
 
2016
;
89
:
1331
1336
58
Scherberich
JE.
The king’s new cloths—the infinite history of kidney tissue proteinuria (histuria) as a renal biomarker
.
Nieren- und Hochdruckkrankh
 
2012
;
41
:
436
450
59
Scherberich
JE
Wolf
G.
Disintegration and recovery of kidney membrane proteins: consequence of acute and chronic renal failure
.
Kidney Int Suppl
 
1994
;
47
:
S52
S57
60
Johnson
KR.
Strengths and weaknesses of renal markers as risk factors and surrogate markers
.
Kidney Int
 
2011
;
79
:
1272
1274
61
Filler
G
Huang
SH
Lindsay
RM.
The search for more reliable estimated GFR biomarkers
.
Am J Kidney Dis
 
2016
;
67
:
5
8
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com

Supplementary data

Comments

0 Comments