Serum-soluble urokinase receptor levels do not distinguish focal segmental glomerulosclerosis from other causes of nephrotic syndrome in children

Transcription

1 & 214 International Society of Nephrology clinical investigation see commentary on page 499 Serum-soluble urokinase receptor levels do not distinguish focal segmental glomerulosclerosis from other causes of nephrotic syndrome in children Aditi Sinha 1, Jaya Bajpai 1, Savita Saini 1, Divya Bhatia 1, Aarti Gupta 1, Mamta Puraswani 1, Amit K. Dinda 2, Sanjay K. Agarwal 3, Shailaja Sopory 4, Ravindra M. Pandey 5, Pankaj Hari 1 and Arvind Bagga 1 1 Division of Nephrology, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India; 2 Department of Pathology, All India Institute of Medical Sciences, New Delhi, India; 3 Department of Nephrology, All India Institute of Medical Sciences, New Delhi, India; 4 Pediatric Biology Centre, Translational Health Science and Technology Institute, Government of India, Gurgaon, India and 5 Department of Biostatistics, All India Institute of Medical Sciences, New Delhi, India In this prospective study, we measured serum levels of the soluble urokinase receptor (supar) in pediatric patients with nephrotic syndrome of various etiologies. Mean levels of supar were 3316 pg/ml in 99 patients with steroid-resistant focal segmental glomerulosclerosis and 3253 pg/ml in 117 patients with biopsy-proven minimal change disease, which were similar to that of 138 patients with steroid-sensitive nephrotic syndrome (315 pg/ml) and 83 healthy controls (321 pg/ml). Similar proportions of patients in each group had supar over 3 pg/ml. Compared with controls, supar levels were significantly higher in patients with focal segmental glomerulosclerosis (FSGS) and estimated glomerular filtration rate (egfr) under 3 ml/min per 1.73 m 2 (6365 pg/ml), congenital nephrotic syndrome (4398 pg/ml), and other proteinuric diseases with or without egfr under 3 ml/min per 1.73 m 2 (552 and 3875 pg/ml, respectively; both significant). There were no changes following therapy and during remission. Levels of supar significantly correlated in an inverse manner with egfr (r ¼.36) and directly with C-reactive protein (r ¼.2). The urinary supar-to-creatinine ratio significantly correlated with proteinuria (r ¼.25) in 151 patients and controls. Using generalized estimating equations approach, serum supar significantly correlated with egfr (coefficient ¼ 13.75), age at sampling (2.72), and C-reactive protein (39.85). Thus, serum supar levels in nephrotic syndrome are similar to controls, and do not discriminate between FSGS, minimal change disease, or steroid-responsive illness. Kidney International (214) 85, ; doi:1.138/ki ; published online 15 January 214 KEYWORDS: focal segmental glomerulosclerosis; minimal change disease; soluble urokinase receptor; steroid-resistant nephrotic syndrome Correspondence: Arvind Bagga, Division of Nephrology, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi 1129, India. arvindbagga@hotmail.com Received 3 June 213; revised 16 November 213; accepted 26 November 213; published online 15 January 214 Focal segmental glomerulosclerosis (FSGS) is an important cause of steroid-resistant nephrotic syndrome in children. Clinical and experimental evidence suggests the pathogenic role of circulating permeability factors, including soluble urokinase plasminogen activating receptor (supar). 1,2 Serum supar levels were found to be elevated in Caucasian adults with primary FSGS 3 and in two cohorts of children with FSGS from Europe and the United States. 4 Animal models further showed that membrane-bound upar activates podocyte b3 integrin signaling, resulting in foot process effacement and a glomerulopathy resembling FSGS. 3,5 However, recent observations have questioned the utility of supar in mediating proteinuria in childhood-onset FSGS and other proteinuric kidney diseases across diverse populations. 6,7 This prospective study examined blood levels of supar in a large, carefully phenotyped cohort of Indian children with steroid-resistant FSGS and minimal change disease, steroidsensitive nephrotic syndrome, other proteinuric chronic kidney diseases (CKDs), and healthy controls. Sequential specimens were collected in a subgroup of patients to examine the relationship of serum supar during nephrotic-range proteinuria and remission. Urinary levels of supar were estimated in a proportion of patients with steroid-resistant FSGS, minimal change disease, and controls. RESULTS During April 212 to May 213, we collected 617 blood samples from 469 patients and 83 controls. Baseline characteristics of patients are shown in Table 1. Patients with FSGS were older than those with steroid-resistant minimal change disease (P ¼.5), similar in age to patients with steroidsensitive nephrotic syndrome and healthy controls, and younger than other proteinuric CKD (Po.1). The estimated glomerular filtration rate (egfr) was below 3 ml/min per 1.73 m 2 in 21 patients with steroid-resistant FSGS, 38 with other proteinuric kidney diseases, and 2 with congenital nephrotic syndrome. Kidney International (214) 85,

2 clinical investigation A Sinha et al.: Serum supar levels in nephrotic syndrome Table 1 Characteristics of patients (n ¼ 469) and healthy controls (n ¼ 83) Steroid-resistant nephrotic syndrome FSGS (n ¼ 12) MCD (n ¼ 117) Steroid-sensitive nephrotic syndrome (n ¼ 138) Congenital nephrotic syndrome (n ¼ 9) Proteinuric chronic kidney disease (n ¼ 85) Controls (n ¼ 83) Boys 83 (69.2%)^^ 85 (72.7%)^^ 113 (81.9%)* 7 (77.8%) 63 (74.1%) 42 (5.6%) Age at onset, months 54.3± ±3.9** 42.8±3.9** 1.1±.9** 82.1±48.9*** ( ) ( ) ( ) (.5 1.3) ( ) Age at sampling, months 113.1±57.1 ( ) 93.2±51.8** ( ) 16.8±49. ( ) 12.2±8.*** ( ) 15.±46.3*** ( ) 1.1±49.7 (6. 24) Hypertension 72 (62.1%) 46 (39.3%)** 52 (39.1%)*** 1 (11.1%)* 51 (6.%) 5 (6.%) egfr, ml/min per 1.73 m ±42.9 ( ) 15.8±32.8*** ( ) 15.2±29.6*** ( ) 51.2±34.9* ( ) 59.3±43.1*** ( ) 95.6±25.4 ( ) Serum albumin, g/dl 3.±1.2 (2. 4.) 3.2±1.3 ( ) 3.2±1.3 ( ) 1.5±.4** ( ) 3.6±1.** ( ) 4.8±.3 (4.6 5.) Abbreviations: egfr, estimated glomerular filtration rate; FSGS, focal segmental glomerulosclerosis; MCD, minimal-change disease. Values represent proportions (%) or mean±s.d. (interquartile range). P versus FSGS ¼ *Po.5, **Po.1, ***Po.1; ^^P versus controls ¼ Po.1. Of 237 patients with steroid-resistant nephrotic syndrome secondary to histologically proven FSGS or minimal-change disease, 62 (26.2%) patients were sampled at diagnosis of corticosteroid resistance. Therapies at the time of sampling included calcineurin inhibitors (n ¼ 62), mycophenolate mofetil (MMF; n ¼ 14), or prednisone (n ¼ 78) with or without enalapril; 21 patients with FSGS and CKD stages 4 and 5 were receiving supportive care. Of 34 patients with FSGS screened for mutations in two genes, 2 had heterozygous WT1 mutation associated with the Frasier syndrome (IVS9 þ 4C4T in intron 9 and IVS9 þ 5G4A in exon 9), 3 had heterozygous variation R229Q, and one each had heterozygous NPHS2 mutation at c.89c4t in exon 8 and c.557_56deltaat in exon 5. Of 138 patients with steroid-sensitive nephrotic syndrome, 19 were not receiving any therapy; others were receiving prednisone alone (n ¼ 1), or levamisole (n ¼ 7), MMF (n ¼ 3), or a calcineurin inhibitor (n ¼ 9). Disorders in patients with other proteinuric CKDs were glomerular diseases in 63 and nonglomerular conditions in 22 patients. The former included membranoproliferative glomerulonephritis type 1 (n ¼ 13) or type 2 (n ¼ 8), pauci-immune crescentic glomerulonephritis (n ¼ 7), secondary FSGS (n ¼ 7; associated with reflux nephropathy, crescentic glomerulonephritis, or hypertension), chronic glomerulonephritis (n ¼ 6), Alport syndrome (n ¼ 5), hemolytic uremic syndrome (n ¼ 4), IgA nephropathy (n ¼ 4), and others (n ¼ 9). Nonglomerular conditions were posterior urethral valves (n ¼ 8), neurogenic bladder (n ¼ 5), nephronophthisis (n ¼ 3), solitary kidney (n ¼ 2), and unknown causes (n ¼ 4). Sixty patients had more than one blood sample in different states of illness. Table 2 shows corresponding egfr, urine protein-to-creatinine ratio (Up/Uc), and highly sensitive C-reactive protein (CRP) levels in samples from various categories. Levels of CRP were similar in patients with steroid resistance secondary to FSGS with egfr 43 ml/min per 1.73 m 2, minimal change disease, and steroid-sensitive nephrotic syndrome (supplementary Figure 1; online; analysis of variance (ANOVA), P ¼.67). Values in healthy controls (.9±1.2 mg/l) were lower than for FSGS (2.4±6.3 mg/l; P ¼.33). CRP values were similar within subgroups of the above categories (all ANOVA P4.1). Compared with controls, these values were significantly higher in patients with CKD stages 4 and 5 associated with FSGS (4.±7.2 mg/l; P ¼.3) and other proteinuric diseases (6.6±16.9 mg/l; P ¼.3). supar levels and underlying etiology Figure 1 shows serum supar in samples from various categories (n ¼ 617). The mean±s.e.m. levels were similar in patients with steroid-resistant nephrotic syndrome associated with FSGS (3316.1±11.2 pg/ml) and minimal change disease (3252.9±115.7 pg/ml; P ¼.69). Levels in patients with FSGS were comparable to steroid-sensitive nephrotic syndrome (3149.6±96.8 pg/ml; P ¼.24) and healthy controls (321.1±155.5 pg/ml; P ¼.11), but lower than proteinuric CKD and egfr 43 ml/min per 1.73 m 2 (3875.± 22.4 pg/ml; P ¼.9) or proteinuric CKD and egfr o3 ml/min per 1.73 m 2 (551.6±386. pg/ml; Po.1). Table 3 shows serum supar in samples from subgroups within each clinical category (n ¼ 617). The levels were similar in patients with FSGS with nephrotic-range proteinuria (332.7±137.4 pg/ml) compared with remission (339.9± 15.1 pg/ml, P ¼.96). Patients with FSGS and egfr o3 ml/ min per 1.73 m 2 (6365.1±65.4 pg/ml) showed higher levels compared with FSGS and egfr 43 ml/min per 1.73 m 2 (3316.1±11.2 pg/ml, P ¼.3), and healthy controls (Po.1). Levels of supar were similar in patients with steroid-resistant minimal change disease during nephroticrange proteinuria (333.7±186.6 pg/ml) and remission (323.6±139.6 pg/ml; P ¼.67). Values were significantly higher in patients with congenital nephrotic syndrome compared with controls (P ¼.9). Patients with steroidsensitive nephrotic syndrome sampled at the initial episode, remission, or relapse showed similar supar levels (ANOVA, P ¼.45) that were also similar to controls (ANOVA, P ¼.42). 65 Kidney International (214) 85,

3 A Sinha et al.: Serum supar levels in nephrotic syndrome clinical investigation Table 2 egfr, highly sensitive CRP, and urine protein-to-creatinine ratio in clinical categories egfr, ml/min per 1.73 m 2 CRP, mg/l Spot urine protein:creatinine, mg/mg Steroid-resistant nephrotic syndrome due to MCD MCD, nephrotic (54) 111.5± ±3.7 3.±2.2 MCD, nonresponse (17) 12.1± ±2. 3.3±3.1 MCD, remission (71) 11.8± ±2.5.2±.4 Steroid-resistant nephrotic syndrome due to FSGS FSGS, nephrotic (46) 15.3± ±8.8 4.±3.8 FSGS, nonresponse (28) 84.8± ± ±6.9 FSGS, remission (52) 97.± ±4.3.3±.5 FSGS with CKD stages 4 and 5 (21) 14.3±8.1 4.±7.2^^ 3.1±1.3 Congenital nephrotic syndrome (9) 51.2± ±2.9^^ 8.3±8.9 Steroid-sensitive nephrotic syndrome Initial episode (19) 15.1± ±3.8^ 4.3±2.7 Relapse (6) 112.8± ± ±1.9 Remission (72) 98.2± ±1.6.1±.3 Proteinuric CKD Glomerular CKD 1 3 (41) 94.4± ± ±4.7 Glomerular CKD 4 and 5 (22) 19.1± ±21.4^^ 2.7±2.2 Nonglomerular CKD 1 3 (6) 55.8± ±1.4.9±.8 Non glomerular CKD 4 and 5 (16) 22.9± ±5.2^^.9±.8 Controls (83) 95.6±25.4.9±1.2.5±.6 Abbreviations: ANOVA, analysis of variance; CKD, chronic kidney disease; CRP, C-reactive protein; egfr, estimated glomerular filtration rate; FSGS, focal segmental glomerulosclerosis; MCD, minimal-change disease. For CRP, P versus controls ¼ ^Po.5, ^^Po.1. ANOVA test did not show significant differences in CRP among subgroups of patients with steroid-resistant MCD, FSGS, and steroid-sensitive nephrotic syndrome. 15, 1, ±11.2 Steroid-resistant FSGS, egfr >3 ml/min *** ** *** ±65.4 (***) Steroid-resistant FSGS, egfr <3 ml/min *** *** ** * *** 551.6±386. (***) 3875.± ±115.7 (**) ±96.8 Steroid-resistant minimal change disease Steroid-sensitive nephrotic syndrome Proteinuric chronic kidney disease, egfr >3 ml/min Proteinuric chronic kidney disease, egfr <3 ml/min 321.1±155.5 Controls Figure 1 Serum soluble urokinase receptor (supar) levels in various clinical categories (n ¼ 617). P-values for differences in mean±s.e.m. values compared with controls are indicated in parentheses, and those between groups above horizontal bars; *Po.5, **Po.1, and ***Po.1. Values for supar in other causes of proteinuria and CKD stages 1 3 (n ¼ 47; 3875.±22.4 pg/ml) were higher than in controls (Po.1) and similar for glomerular and nonglomerular causes (P ¼.6). Levels of supar in seven patients with secondary FSGS and egfr 71.±24.3 ml/min per 1.73 m 2 tended to be higher (417.5±32.9 pg/ml) compared with primary FSGS (3316.1±11.2 pg/ml; P ¼.98). Values in patients with proteinuric CKD stages 4 and 5 (n ¼ 38; 551.6±386. pg/ml) were higher than CKD stages 1 3 (P ¼.7) or healthy controls (Po.1). Among samples from patients with FSGS, 58.7% had supar levels that exceeded 3 pg/ml. The proportion was similar in patients with steroid-resistant minimal change disease (51.4%; P ¼.24), steroid-sensitive nephrotic syndrome (49.7%; P ¼.14), congenital nephrotic syndrome (66.7%; P ¼.67), and other proteinuric CKD with egfr 43 ml/ min per 1.73 m 2 (68.1%; P ¼.3). Compared with controls, a significantly higher proportion of samples from patients with egfr o3 ml/min per 1.73 m 2, including 95.2% with FSGS (P ¼.1) and 78.9% with proteinuric CKD (P ¼.28), showed supar 43 pg/ml. Levels of supar were comparable for girls and boys in the entire group (3429.6±14.5 vs ±86.6 pg/ml; P ¼.56) and in patients with FSGS (3441.4±174.4 vs ± pg/ml; P ¼.17), steroid-resistant minimal change Kidney International (214) 85,

4 clinical investigation A Sinha et al.: Serum supar levels in nephrotic syndrome Table 3 Serum and urinary levels of supar in samples from patients and controls Disease (n), mean±s.e.m. Proportion with supar43 pg/ml Urinary supar:creatinine, ng/g, mean±s.d. (number tested) Steroid-resistant nephrotic syndrome due to MCD a,b MCD, nephrotic (54) 342.7± % 9.8±5.3 (6) MCD, nonresponse (17) ± % 15.5±12. (2) MCD, remission (71) 323.6± % 3.8±1. (2) Steroid-resistant nephrotic syndrome due to FSGS b FSGS, nephrotic (46) 314.3± % 12.8±6.8 (25) FSGS, nonresponse (28) ±259.8* 6.7% 4.7±5.9 (2) FSGS, remission (52) 339.9± % 2.9±1.4 (6) FSGS and chronic kidney disease (CKD) 4 and 5 (21) ±65.4 ###, *** 95.2% ###, *** 13.6±9.6 (1) Congenital nephrotic syndrome (9) ±978.5 #, * 66.7% ND Steroid-sensitive nephrotic syndrome Initial episode (19) 3343.± % ND Relapse (6) 35.4± % 8.9±5.5 (27) Remission (72) ± % 9.9±13.9 (23) Proteinuric CKD a,b,c Glomerular CKD 1 3 (41) 383.5±238.3** 68.3%* 7.±4.3 (13) Glomerular CKD 4 and 5 (22) 524.8±538.** 77.3%** 4.2±2.3 (11) Nonglomerular CKD 1 3 (6) ± % ND Nonglomerular CKD 4 and 5 (16) 484.8±557.5** 81.3%** 3.6±.8 (3) Controls (83) 321.1± % 8.9±8.3 (12) Abbreviations: ANOVA, analysis of variance; CKD, chronic kidney disease; CRP, C-reactive protein; egfr, estimated glomerular filtration rate; FSGS, focal segmental glomerulosclerosis; MCD, minimal-change disease; ND, not done; supar, soluble urokinase receptor. P versus controls ¼ *Po.5, **Po.1, ***Po.1; P versus FSGS ¼ # Po.5, ## Po.1, ### Po.1. Significant within group differences were seen on ANOVA test for a proportion having supar 43 pg/ml, b urinary supar/creatinine ratio, and c serum supar level. disease (343.3±273.1 vs ±143.3 pg/ml; P ¼.45), steroid-sensitive nephrotic syndrome (3184.7±156.3 vs ±118.6 pg/ml; P ¼.71), and controls (313.9±141.4 vs ±131.9 pg/ml; P ¼.31). Patients with hypertension showed higher levels than those without (3356.±89.8 vs ±129.7 pg/ml; P ¼.4). In patients with steroid resistance, values were comparable in patients with (n ¼ 39) or without (n ¼ 199) tubulointerstitial fibrosis (3787.± 29. vs ±123.5 pg/ml; P ¼.35). Levels were similar in initial (3264.4±122. pg/ml) or late steroid resistance (3319.7±12.1 pg/ml, P ¼.75). Serum supar levels correlate with renal function and CRP Serum supar correlated inversely with egfr in the entire cohort (r ¼.36, Po.1) and in patients with FSGS (r ¼.51, Po.1; Figure 2a and b). Receiver operating characteristic analyses for egfro3 ml/min per 1.73 m 2 showed that supar values of and pg/ml correctly classified 85.1% of the entire cohort and 88.1% of patients with FSGS, respectively (Figure 3). A direct correlation was observed between supar and CRP in the entire group (r ¼.2, Po.1) and patients with FSGS (r ¼.27, P ¼.3; Figure 2c and d). Values of CRP correlated inversely with egfr (r ¼.11, P ¼.1). Serum supar levels do not relate to proteinuria or immunosuppressive therapy Levels of supar were similar among patients with steroidresistant nephrotic syndrome, due to FSGS and minimal change disease, in remission (n ¼ 81; ±135.1 pg/ml), non-nephrotic-range proteinuria (n ¼ 29; ±196.8 pg/ml), and nephrotic-range proteinuria (n ¼ 127; 37.1±16.9 pg/ml; ANOVA, P ¼.32). Values of supar did not correlate with proteinuria, estimated by Up/Uc ratio, in the overall patient cohort (r ¼.2, P ¼.7) or in patients with FSGS (r ¼.12, P ¼.29). Archived sera, from patients with FSGS, who underwent transplantation, showed no difference in pretransplant supar levels in recurrent FSGS (n ¼ 1) compared with six with no recurrence ( vs ± pg/ml; P ¼.32). The mean duration of therapy with calcineurin inhibitors and MMF for steroid-resistant nephrotic syndrome was 11.9±11.2 and 15.3±13.6 months, respectively, and levels of supar were similar at ±121. and ±48. pg/ml (P ¼.73). There was no difference in levels in patients receiving cyclosporine (n ¼ 2) versus tacrolimus (n ¼ 42; ±124. and 385.2±277.6 pg/ml; P ¼.89). In 6 patients with nephrotic syndrome, 1 (n ¼ 55) or 2(n ¼ 5) additional blood samples were analyzed at least 4 weeks apart. In patients with FSGS, minimal-change 652 Kidney International (214) 85,

5 A Sinha et al.: Serum supar levels in nephrotic syndrome clinical investigation a c 15, 1, 5 15, 1, 5 b egfr, ml/min per 1.73 m 2 egfr, ml/min per 1.73 m 2 Highly sensitive C-reactive protein, mg/l d 15, 1, 5 15, 1, Highly sensitive C-reactive protein, mg/l Figure 2 Serum levels of soluble urokinase receptor (supar; pg/ml) in relation to estimated glomerular filtration rate (egfr) and highly sensitive C-reactive protein (CRP). An inverse relationship was seen between serum supar and egfr in (a) all patients and controls (n ¼ 552; Po.1), and (b) patients with focal segmental glomerulosclerosis (FSGS; n ¼ 12; Po.1). A direct correlation was noted between levels of supar and CRP in (c) all patients and controls (n ¼ 552; Po.1), and (d) patients with FSGS (n ¼ 12; P ¼.3). disease, and steroid-sensitive illness, supar levels were similar at baseline and following complete or partial remission (Figure 4a c). Although values did not increase in patients showing nonresponse to therapy with calcineurin inhibitors (Figure 4d), these were insignificantly higher in patients showing egfr o3 ml/min per 1.73 m 2 (Figure 4e). Urinary supar excretion correlates with proteinuria Urinary supar, expressed as the ratio to urinary creatinine (ng/g) in 16 samples from 151 patients, did not correlate with its serum levels in the entire group (r ¼.14; P ¼.96) nor in patients with FSGS (r ¼.14, P ¼.34; Figure 5). Urinary levels were inversely correlated with age (r ¼.19, P ¼.2), weight (r ¼.23, P ¼.7), and height (r ¼.24, P ¼.2), and directly to the Up/Uc ratio (r ¼.25, P ¼.7) in the entire group. Similarly, the levels were related to height (r ¼.28, P ¼.47) and Up/Uc ratio (r ¼.34, P ¼.4) in patients with FSGS. These associations were confirmed when examined using the generalized estimating equations (GEE) approach (Table 4). Further, urinary supar excretion was independently related to the Up/Uc ratio (b-coefficient±s.e.m. ¼.43±.24; P ¼.49). Correlation of serum supar with patient characteristics Serum levels of supar correlated with age at sampling (r ¼.12, P ¼.5), egfr (r ¼.36, Po.1), and CRP (r ¼.2, Po.1). The respective correlations persisted in patients with FSGS (r ¼.24, P ¼.8; r ¼.51, Po.1; and r ¼.27, P ¼.26). These associations were confirmed when examined using the GEE approach (Table 4). The GEE approach showed that levels of supar were predicted using egfr (b-coefficient ¼ 13.75±1.7; Po.1), age at sampling (2.72±1.22; P ¼.26), and CRP (39.85±1.92; Po.1) with a constant of ± 225. pg/ml. For FSGS, the coefficients were as follows: egfr ¼ 23.99±3.56 (Po.1), age at sampling ¼ 5.8± 2.6 (P ¼.53), and CRP ¼ 85.68±2.39 (Po.1), with a constant of ± pg/ml. DISCUSSION This study, on almost 4 children with nephrotic syndrome, shows that levels of serum supar were similar in patients with steroid-resistant and steroid-sensitive nephrotic syndrome, did not vary between disease relapse and remission, and failed to discriminate FSGS from minimal-change disease. Further, the levels in patients with nephrotic syndrome were comparable to those in healthy controls and other proteinuric kidney diseases. Regardless of cause, these values correlated inversely with egfr and directly with CRP. Approximately one-half of all patients with steroidresistant and steroid-sensitive nephrotic syndrome and 85% of those with CKD stages 4 and 5 showed supar above 3 pg/ml. Urinary supar was directly related to the level of proteinuria. The finding that blood levels of supar were elevated in patients with recurrent FSGS has revived interest on the elusive circulating factor for FSGS. 3 Membrane-bound upar on podocytes is believed to cause proteinuria by activation of b3-integrin signaling, 3,5 and this axis might be downregulated by calcineurin inhibitors, 8 amiloride, 9 and calcitriol. 1 Kidney International (214) 85,

6 clinical investigation A Sinha et al.: Serum supar levels in nephrotic syndrome Sensitivity Sensitivity Entire cohort.5 1 specificity Patients with FSGS specificity 1. Figure 3 Receiver operating characteristics (ROCs). ROCs to demonstrate the relationship between estimated glomerular filtration rate (egfr) below 3 ml/min per 1.73 m 2 and serum levels of soluble urokinase receptor (supar) in (a) all patients and controls (n ¼ 552), and (b) patients with focal segmental glomerulosclerosis (FSGS; n ¼ 12). Values of supar predictive of egfro3 ml/min per 1.73 m 2 were pg/ml for the entire group (correctly classified 85.1%; sensitivity 74.5%, specificity 83.7%; area under the curve ¼.81) and pg/ml for patients with FSGS (correctly classified, 88.1%; sensitivity, 88.2%, specificity, 84.2%; area under the curve ¼.93). Other workers, using the same enzyme-linked immunosorbent assay kit, have examined the association between circulating supar and FSGS with conflicting results. 4,6,7,11,12 Levels of supar were higher than age-matched controls in 7 patients with FSGS from the NIH-FSGS trial and 94 patients from the PodoNet registry. 4 The levels were elevated above 3 pg/ml in 84.3% and 55.3% patients, respectively, compared with 6% controls; the increased levels were not attributed to an inflammatory state. Blood levels declined following therapy with MMF, but increased following treatment with calcineurin inhibitors. 4 However, Maas et al. 6 found no differences in blood levels of supar in a small group (n ¼ 23) of patients with primary and secondary FSGS and minimal-change disease. Similarly, supar failed to distinguish FSGS from other etiologies when evaluating pretransplant sera from patients with advanced renal disease. 11 In another report, serum supar, although higher in adult patients, failed to distinguish between primary and secondary forms of FSGS. 12 In a recent singlecenter study on 99 patients, levels of supar were similar between patients with FSGS, non-fsgs glomerular disease, and healthy controls. 7 Our findings support these results in a larger, younger, and phenotypically diverse cohort of patients. Serum supar, as measured by current enzyme immunoassays, does not distinguish FSGS from minimal change disease, steroid-sensitive nephrotic syndrome, and other causes of glomerular proteinuria. These levels did not decline following prolonged therapy, either with MMF or calcineurin inhibitors. Levels of supar are elevated in a number of conditions including sepsis, 13,14 malaria, 15 and chronic infections including HIV, 16 hepatitis C, 17 and tuberculosis. 18 However, these conditions are not frequently associated with nephroticrange proteinuria, casting doubt on the hypothesis that relates supar levels to podocyte injury and proteinuria. High levels of supar, tumor necrosis factor-a, and CRP are reported in patients with CKD 19,2 and in those receiving hemodialysis, 21 suggesting a state of inflammation. A proportion of patients with FSGS and other proteinuric diseases, in the present study, had increased levels of CRP. We also found that levels of supar were directly correlated to CRP, suggesting that inflammation-induced synthesis might contribute to elevated levels of supar The finding that serum supar levels are increased in patients with reduced GFR is not novel 21 and perhaps reflect its impaired excretion. Studies in patients with nephrotic syndrome show an inverse relationship between serum supar and egfr, 4,6,7,11,12 which were confirmed in the present report. Serum levels of supar, therefore, need to be evaluated in context of renal function and/or the presence of systemic inflammation. Our findings that blood levels of supar were similar in boys and girls are contrary to previous reports that showed higher 7,16,22 or lower 4 mean values of supar in girls compared with boys. The relationship of serum supar with age is also not clear. Although some reports suggest that age was inversely related to serum supar, 7 the present and few other studies show that supar levels increased with age. 16,23 These results emphasize the need to ensure adequate number of age- and sex-matched controls in future studies involving children. Urinary excretion of upar is considered a biomarker in urosepsis, 24 renal allograft rejection, 25 breast cancer, 26 and hematological malignancies, 27 but has similar or weaker 18 association with disease severity or mortality than serum supar. A recent report in 86 renal transplant recipients showed that levels of urine, and not serum supar, were higher in patients with recurrent FSGS. 11 The study also showed that urinary upar levels correlated with serum supar and urinary protein excretion, but not with egfr. Studies in healthy volunteers 28 and HIV-infected patients 23 show a linear correlation between levels of urinary and serum supar. The present study confirms the association of supar excretion with proteinuria, but not with serum supar. This study has some limitations. Except for a few, most of the blood specimens were not taken at the onset of the illness but at different stages of the disease. Most patients had only one sample, limiting the ability to study supar levels longitudinally. Although we did not find that therapy or 654 Kidney International (214) 85,

7 A Sinha et al.: Serum supar levels in nephrotic syndrome clinical investigation a 6 b At onset or relapse During remission Focal segmental glomerulosclerosis At onset or relapse During remission Minimal change disease c d e 1, At onset or relapse During remission Steroid-sensitive nephrotic syndrome At onset or relapse Non response Focal segmental glomerulosclerosis Onset or non response egfr <3 ml/ min per 1.73 m 2 Focal segmental glomerulosclerosis Figure 4 Blood levels of soluble urokinase receptor (supar) in patients with paired samples. (a) Steroid-resistant nephrotic syndrome due to focal segmental glomerulosclerosis (FSGS; n ¼ 18): supar levels were ±226.5 pg/ml at diagnosis or relapse, and ±246.8 pg/ml during therapy-induced complete or partial remission (P ¼.7). (b) Steroid-resistant nephrotic syndrome due to minimal-change disease (n ¼ 18): supar levels were ±184.9 pg/ml at diagnosis or relapse, and ±217.7 pg/ml during complete or partial remission (P ¼.96). (c) Steroid-sensitive nephrotic syndrome (n ¼ 14): supar levels were ±544.1 pg/ml at the onset of illness (n ¼ 3) or at relapse (n ¼ 11), and ±25.8 pg/ml following corticosteroid-induced remission (P ¼.73). (d) FSGS with nonresponse (n ¼ 5): supar levels were 294.3±294.7 pg/ml at baseline and ±576.6 pg/ml following 6 months of therapy with tacrolimus (P ¼.27). (e) FSGS with progressive impairment of function (n ¼ 5): supar levels increased from ±343. pg/ml at baseline to ±155.2 pg/ml in patients with estimated glomerular filtration rate (egfr)o3 ml/min per 1.73 m 2 (P ¼.31). 8 8 Urinary supar to creatinine, ng/g Urinary supar to creatinine, ng/g 5 1, 15, 5 1, 15, Urinary supar to creatinine, ng/g Urinary supar to creatinine, ng/g Urinary protein to creatinine, mg/mg Urinary protein to creatinine, mg/mg Figure 5 Levels of urinary soluble urokinase receptor (supar)-to-creatinine ratio (ng/g) in relation to serum supar and proteinuria. There was no correlation of urinary supar and serum supar in (a) all patients and controls (n ¼ 151; P ¼.96), and (b) patients with focal segmental glomerulosclerosis (FSGS; n ¼ 5; P ¼.34). A direct correlation was noted between values of urinary supar and urinary protein-tocreatinine ratio in samples from (c) all patients and controls (n ¼ 151; P ¼.7), and (d) patients with FSGS (n ¼ 5; P ¼.4) Kidney International (214) 85,

8 clinical investigation A Sinha et al.: Serum supar levels in nephrotic syndrome Table 4 Association of patient characteristics with serum and urinary supar in samples from all patients and controls (entire cohort) and patients with FSGS Urinary supar to creatinine, ng/g Entire cohort, n ¼ 617 Patients with FSGS, n ¼ 147 Entire cohort, n ¼ 16 Patients with FSGS, n ¼ 61 Coefficient±s.e.m. P Coefficient±s.e.m. P Coefficient±s.e.m. P Coefficient±s.e.m. P Male sex 86.87± ± ± ± Age at sampling, months 3.63± ± ± ± Duration of disease, months 1.57± ± ± ± Presence of hypertension 466.2± ± ± ± Height, cm 3.74± ± ±.3.1.9±.4.21 Weight, kg 3.97± ± ± ±.9.62 Body mass index 34.9± ± ± ±.4.72 egfr, ml/min per 1.73 m 2 15.±1.63 o ±3.37 o.1.5 ± ± Spot urine protein: 9.56± ± ± ±.17.4 creatinine, mg/mg Highly sensitive CRP, mg/l 53.46±11.15 o ± ± ± Urinary supar: creatinine, ng/g 25.74± ± ± ± Abbreviations: CRP, C-reactive protein; egfr, estimated glomerular filtration rate; FSGS, focal segmental glomerulosclerosis; supar, soluble urokinase receptor. occurrence of remission altered supar levels, the effect of prolonged illness and medications on renal functions needs closer examination. Although we found that o1% patients with FSGS have mutations in WT1 or NPHS2 (unpublished data), a detailed genetic evaluation of the present patients was desirable but could not be conducted. Only one patient with recurrent FSGS, reported to show the highest levels of supar, 3,11 was examined in this report. Finally, compared with previous reports that mean levels in controls range between 1739 and 274 pg/ml, 7,12 mean supar levels in healthy children in this study were pg/ml. The reasons for high normal values are unclear, particularly as conditions that increase supar such as renal or hepatic disease, malnutrition, and acute or chronic inflammation were excluded. However, others have reported higher levels in healthy children and adults, with values ranging between 31 and 394 pg/ml. 15,29,3 Despite these concerns, this prospective study on a large number of patients and controls shows limited clinical utility of measuring serum supar in children with steroid-resistant nephrotic syndrome. Serum levels of supar do not discriminate between patients with FSGS and minimal-change disease, or steroid-resistant and steroid-sensitive nephrotic syndrome. The levels of supar were inversely correlated with renal function and directly with evidence of systemic inflammation. Our results fail to support published literature that serum supar, as measured by current immunoassays, is the circulating factor that results in FSGS and emphasize the need for studies to examine the role of biomarkers in the pathogenesis of nephrotic syndrome. MATERIALS AND METHODS Following approval by the Ethics Committee and parental consent, we screened all consecutive patients with idiopathic steroid-resistant nephrotic syndrome attending the hospital between April 212 and April 213. Patients, 1 18 years old, with initial or late steroid resistance and histological features suggestive of FSGS or minimalchange disease, were eligible. Initial resistance was the absence of remission despite therapy with prednisone at 2 mg/kg/day (maximum 6 mg) for 4 weeks; patients with remission at onset but steroid resistance during a subsequent relapse were defined as having late resistance. Patients with nephrotic-range proteinuria (Up/Uc42; dipstick 3 4 þ ) at diagnosis of resistance or during relapse were categorized separately from those with complete or partial remission, nonresponse to therapy, and impaired renal function (egfr o3 ml/min per 1.73 m 2 ). Blood levels of supar were also estimated in patients with the following: (i) steroid-sensitive nephrotic syndrome at onset, disease relapse, or during remission; (ii) congenital nephrotic syndrome; and (iii) other CKDs with proteinuria (Up/Uc4.2; dipstick 2 þ or more). Patients with other proteinuric CKDs were categorized into glomerular or nonglomerular disease and by egfr (CKD stages 1 3 or stages 4 5). 31 Paired sera were obtained, at least 28 days apart, in a subset of patients with steroid-sensitive or steroid-resistant nephrotic syndrome during remission or subsequent relapse. Controls were children, aged 1 18 years, presenting for minor surgeries, evaluation of noninflammatory illnesses, or vaccination. Subjects with recent (o4 weeks) bacterial infection, diarrhea or malaria, chronic infections, malnutrition, and known metabolic, hepatic, or renal diseases were excluded. Clinical details Diagnoses and treatment were confirmed by review of medical records. Spot Up/Uc ratio and blood levels of creatinine, albumin, and cholesterol were estimated. Standard definitions were used for hypertension 32 and staging of CKD. 33 Renal biopsy specimens were evaluated by light and immunofluorescence microscopy. 34 Patients with the initial episode of nephrotic syndrome were treated with daily prednisolone at 2 mg/kg/day for 6 weeks and 1.5 mg/kg on alternate days for 6 weeks. Relapses were treated with daily prednisolone until remission, followed by alternate days for 4 weeks. 35 Patients with steroid-resistant nephrotic syndrome were treated with a calcineurin inhibitor (tacrolimus or cyclosporine) titrated by trough levels, enalapril and tapering alternate day prednisone. Response, after therapy for 6 months, was defined as complete 656 Kidney International (214) 85,

9 A Sinha et al.: Serum supar levels in nephrotic syndrome clinical investigation remission (Up/Uco.2, serum albumin43.5 g/dl), partial remission (Up/Uc ¼.2 2., serum albumin ¼ g/dl), or nonresponse (Up/Uc42., serum albumin o2.5 g/dl). 36 Serum and urinary supar and serum CRP Serum was separated within 3 min of blood draw and stored at 8 1C in multiple aliquots until analyzed. Supernatant from fresh midstream urine was stored at 8 1C. Specimens were thawed once to estimate serum and urine supar and highly sensitive CRP. supar was measured using the Quantikine Human upar Immunoassay (R&D Systems, Minneapolis, MN). 3 Briefly, 1 ml of assay diluent was added to each well of the polystyrene microplate, precoated with mouse monoclonal antibody, followed by 5 ml of serum or urine, and incubated for 2 h at room temperature. Subsequently, upar conjugate was added to each well, incubated for 2 h, followed by the addition of substrate solution, and the optical density was read at 45 nm. The intra-assay and interassay coefficients of variation were 2.9% and 2.1%, respectively. Urinary supar was expressed as a ratio to creatinine (ng/g). Serum CRP was measured using high-sensitivity enzyme immunoassay (BioCheck, Foster City, CA). Undiluted standards (1 ml) and 1-fold-diluted sera and positive controls (9 ml each) were dispensed in wells of microtiter plate precoated with mouse monoclonal antibody, followed by the addition of enzyme conjugate and substrate solution, and optical density was read at 45 nm. The intra-assay and interassay coefficients of variation were 9.3% and 5.1%, respectively. Statistical analyses Data were analyzed using Stata version 11.2 (StataCorp 29; Stata Statistical Software, College Station, TX) and expressed as mean±s.d. (interquartile range), except mean±s.e.m. for levels of supar. Tests for significance included w 2 -test, paired or unpaired t-tests, and ANOVA; Po.5 was considered significant. In view of repeated samples, we used the GEE approach with linear link and unstructured correlation to evaluate the association between serum or urinary supar and variables of interest. Linear relationships were examined using the Pearson s correlation coefficient. Receiver operating characteristic analyses were performed for serum supar for egfr o3 ml/min per 1.73 m 2 as the dichotomous variable. Repeat samples were excluded from correlation and receiver operating characteristic analyses. On the basis of the method proposed by Altman 37 and the s.d. for serum supar in patients with FSGS (16.9 pg/ml), the study had 68% power to detect a mean difference of 5 pg/ml (standardized difference.497) at a significance of.5 between patients with FSGS (n ¼ 99) and healthy controls (n ¼ 83). DISCLOSURE All the authors declared no competing interest. ACKNOWLEDGMENTS A research grant was provided by the Department of Biotechnology, Government of India BT/PR4629/MED/3/814/212. This study was supported by funding from the Glue Grant Scheme, Department of Biotechnology, Government of India. We thank Uma Kanga and Sanjeev Goswami, Department of Transplant Immunology, All India Institute of Medical Sciences, New Delhi; Satyajit Rath, National Institute of Immunology, New Delhi; and Friedhelm Hildebrandt, Department of Medicine, Howard Hughes Medical Institute, Boston Children s Hospital, Boston, USA. SUPPLEMENTARY MATERIAL Figure S1. Values of serum C-reactive protein in various clinical categories (n ¼ 617). Supplementary material is linked to the online version of the paper at REFERENCES 1. Savin VJ, Sharma R, Sharma M et al. Circulating factor associated with increased glomerular permeability to albumin in recurrent focal segmental glomerulosclerosis. N Engl J Med 1996; 334: Parikh SM. Circulating mediators of focal segmental glomerulosclerosis: soluble urokinase plasminogen activator receptor in context. Am J Kidney Dis 212; 59: Wei C, El Hindi S, Li J et al. Circulating urokinase receptor as a cause of focal segmental glomerulosclerosis. Nat Med 211; 17: Wei C, Trachtman H, Li J et al. Circulating supar in two cohorts of primary FSGS. J Am Soc Nephrol 212; 23: Wei C, Möller CC, Altintas MM et al. Modification of kidney barrier function by the urokinase receptor. Nat Med 28; 14: Maas RJ, Wetzels JF, Deegens JK. Serum-soluble urokinase receptor concentration in primary FSGS. Kidney Int 212; 81: Bock ME, Price HE, Gallon L et al. Serum soluble urokinase-type plasminogen activator receptor levels and idiopathic FSGS in children: a single-center report. Clin J Am Soc Nephrol 213; 8: Zhang B, Shi W, Ma J et al. The calcineurin-nfat pathway allows for urokinase receptor-mediated beta3 integrin signaling to cause podocyte injury. J Mol Med (Berl) 212; 9: Zhang B, Xie S, Shi W et al. Amiloride off-target effect inhibits podocyte urokinase receptor expression and reduces proteinuria. Nephrol Dial Transplant 212; 27: Ma J, Zhang B, Liu S et al. 1,25-dihydroxyvitamin D(3) inhibits podocyte upar expression and reduces proteinuria. PLoS One 213; 8: e Franco Palacios CR, Lieske JC, Wadei HM et al. Urine but not serum soluble urokinase receptor (supar) may identify cases of recurrent FSGS in kidney transplant candidates. Transplantation 213; 96: Huang J, Liu G, Zhang YM et al. Plasma soluble urokinase receptor levels are increased but do not distinguish primary from secondary focal segmental glomerulosclerosis. Kidney Int 213; 84: Koch A, Voigt S, Kruschinski C et al. Circulating soluble urokinase plasminogen activator receptor is stably elevated during the first week of treatment in the intensive care unit and predicts mortality in critically ill patients. Crit Care 211; 15: R Giamarellos-Bourboulis EJ, Norrby-Teglund A, Mylona V et al. Risk assessment in sepsis: a new prognostication rule by APACHE II score and serum soluble urokinase plasminogen activator receptor. Crit Care 212; 16: R Ostrowski SR, Ullum H, Goka BQ et al. Plasma concentrations of soluble urokinase-type plasminogen activator receptor are increased in patients with malaria and are associated with a poor clinical or a fatal outcome. J Infect Dis 25; 191: Langkilde A, Petersen J, Klausen HH et al. Inflammation in HIV-infected patients: impact of HIV, lifestyle, body composition and demography: A cross sectional cohort study. PLoS One 212; 7: e Berres ML, Schlosser B, Berg T et al. Soluble urokinase plasminogen activator receptor is associated with progressive liver fibrosis in hepatitis C infection. J Clin Gastroenterol 212; 46: Rabna P, Andersen A, Wejse C et al. Urine supar levels compared with plasma supar levels as predictors of post-consultation mortality risk among individuals assumed to be TB-negative: a prospective cohort study. Inflammation 21; 33: Pawlak K, Buraczewska-Buczko A, Mysliwiec M et al. Hyperfibrinolysis, upa/supar system, kynurenines, and the prevalence of cardiovascular disease in patients with chronic renal failure on conservative treatment. Am J Med Sci 21; 339: Yeo ES, Hwang JY, Park JE et al. Tumor necrosis factor (TNF-alpha) and C-reactive protein (CRP) are positively associated with the risk of chronic kidney disease in patients with type 2 diabetes. Yonsei Med J 21; 51: Almroth G, Lönn J, Uhlin F et al. Fibroblast growth factor 23, hepatocyte growth factor, interleukin-6, high sensitivity c-reactive protein and soluble urokinase plasminogen activator receptor. Inflammation markers in chronic hemodialysis patients? Scand J Immunol 213; 78: Kidney International (214) 85,

10 clinical investigation A Sinha et al.: Serum supar levels in nephrotic syndrome 22. Stephens RW, Pedersen A, Nielsen HJ et al. ELISA determination of soluble urokinase receptor in blood from healthy donors and cancer patients. Clin Chem 1997; 43: Andersen O, Eugen-Olsen J, Kofoed K et al. Soluble urokinase plasminogen activator receptor is a marker of dysmetabolism in HIVinfected patients receiving highly active antiretroviral therapy. J Med Virol 28; 8: Florquin S, van den Berg JG, Olszyna DP et al. Release of urokinase plasminogen activator receptor during urosepsis and endotoxemia. Kidney Int 21; 59: Roelofs JJ, Rowshani AT, van den Berg JG et al. Expression of urokinase plasminogen activator and its receptor during acute renal allograft rejection. Kidney Int 23; 64: Soydinc HO, Duranyildiz D, Guney N et al. Utility of serum and urine upar levels for diagnosis of breast cancer. Asian Pac J Cancer Prev 212; 13: Mustjoki S, Sidenius N, Sier CF et al. Soluble urokinase receptor levels correlate with number of circulating tumor cells in acute myeloid leukemia and decrease rapidly during chemotherapy. Cancer Res 2; 6: Sier CF, Sidenius N, Mariani A et al. Presence of urokinase-type plasminogen activator receptor in urine of cancer patients and its possible clinical relevance. Lab Invest 1999; 79: Intzilakis T, Hartmann G, Mouridsen MR et al. Soluble urokinase plasminogen activator receptor, C-reactive protein and triglyceride are associated with heart rate variability in non-diabetic Danes. Eur J Clin Invest 213; 43: Enocsson H, Wetterö J, Skogh T et al. Soluble urokinase plasminogen activator receptor levels reflect organ damage in systemic lupus erythematosus. Transl Res 213; 162: Schwartz GJ, Munoz A, Schneider MF et al. New equations to estimate GFR in children with CKD. J Am Soc Nephrol 29; 2: National High Blood Pressure Education Program Working Group on High Blood Pressure in Children and Adolescents. The fourth report on the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents. Pediatrics 24; 114: National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification and stratification. Am J Kidney Dis 22; 39: S1 S International Study of Kidney Disease in Children. Primary nephrotic syndrome in children: clinical significance of histopathologic variants of minimal change and of diffuse mesangial hypercellularity. Kidney Int 1981; 2: Indian Pediatric Nephrology Group, Indian Academy of Pediatrics, Bagga A, Ali U et al. Management of steroid sensitive nephrotic syndrome: revised guidelines. Indian Pediatr 28; 45: Indian Society of Paediatric Nephrology, Bagga A, Gulati A et al. Management of steroid resistant nephrotic syndrome. Indian Pediatr 29; 46: Altman DG.. Statistics and ethics in medical research-iii: How large a sample? Br Med J 198; 281: Kidney International (214) 85,