Lupus Strains and Human Lupus Nephritis Chemokine Ligand 16 in ...

The Journal of Immunology

Elevated Urinary VCAM-1, P-Selectin, Soluble TNF Receptor-1, and CXC Chemokine Ligand 16 in Multiple Murine Lupus Strains and Human Lupus Nephritis1 Tianfu Wu,* Chun Xie,* Hong W. Wang,† Xin J. Zhou,† Noa Schwartz,‡ Sergio Calixto,* Meggan Mackay,§ Cynthia Aranow,§ Chaim Putterman,‡ and Chandra Mohan2* In an effort to identify potential biomarkers in lupus nephritis, urine from mice with spontaneous lupus nephritis was screened for the presence of VCAM-1, P-selectin, TNFR-1, and CXCL16, four molecules that had previously been shown to be elevated in experimental immune nephritis, particularly at the peak of disease. Interestingly, all four molecules were elevated ⬃2- to 4-fold in the urine of several strains of mice with spontaneous lupus nephritis, including the MRL/lpr, NZM2410, and B6.Sle1.lpr strains, correlating well with proteinuria. VCAM-1, P-selectin, TNFR-1, and CXCL16 were enriched in the urine compared with the serum particularly in active disease, and were shown to be expressed within the diseased kidneys. Finally, all four molecules were also elevated in the urine of patients with lupus nephritis, correlating well with urine protein levels and systemic lupus erythematosus disease activity index scores. In particular, urinary VCAM-1 and CXCL16 showed superior specificity and sensitivity in distinguishing subjects with active renal disease from the other systemic lupus erythematosus patients. These studies uncover VCAM-1, P-selectin, TNFR-1, and CXCL16 as a quartet of molecules that may have potential diagnostic significance in lupus nephritis. Longitudinal studies are warranted to establish the clinical use of these potential biomarkers. The Journal of Immunology, 2007, 179: 7166 –7175.

S

ystemic lupus erythematosus (SLE)3 is a systemic autoimmune disease where nephritis accounts for a substantial degree of the morbidity and mortality, both in patients and in mouse models of the disease (1–3). Unfortunately, the clinical management of nephritis in this autoimmune disease is still a difficult problem (3). One thing that is certain is that early diagnosis (and treatment) of lupus nephritis is associated with better disease outcome (4). Hence, it would be very beneficial if one could detect the presence of nephritis early in disease. Although serial renal biopsies may be ideal in closely monitoring the progression of renal disease (5), this may not be the most practical or feasible approach. Clearly, there is an urgent need for a “surrogate” marker of renal disease that one could potentially use to predict the onset of immune nephritis and to monitor its progression in lupus, as emphasized by others (6). *Department of Internal Medicine (Rheumatology) and Center for Immunology, and † Department of Pathology, University of Texas Southwestern Medical School, Dallas, TX 75235; ‡Albert Einstein College of Medicine, Bronx, NY 10461; and §Division of Rheumatology, Columbia University, NY 10027 Received for publication November 29, 2006. Accepted for publication August 29, 2007. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported in part by National Institutes of Health Grants AR44894, AR50812, AR48692, and AI51392, The Alliance for Lupus Research, a Medical Student Research Preceptorship Award from the American College of Rheumatology Research and Education Foundation, a grant from the New York Chapter of the Arthritis Foundation, and the Gina Finzi Memorial Student Summer Fellowship (to N.S.) from the Lupus Foundation of America. 2 Address correspondence and reprint requests to Dr. Chandra Mohan, Department of Internal Medicine/Rheumatology, University of Texas Southwestern Medical Center, Mail Code 8884, Y8.204, 5323 Harry Hines Boulevard, Dallas, TX 75390-8884. E-mail address: [email protected] 3 Abbreviations used in this paper: SLE, systemic lupus erythematosus; SLEDAI, SLE disease activity index; RA, rheumatoid arthritis; ROC, receiver operating characteristic; AUC, area under the curve; DC, dendritic cell; uPAR, urokinase-type plasminogen activator; GBM, glomerular basement membrane; sTNFR-1, soluble TNFR-1.

Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 www.jimmunol.org

Currently, the presence of nephritis is gauged by measuring circulating and excreted indicators of renal dysfunction, with supporting information from renal biopsy. The measurement of excreted urine protein or albumin appears to be the most reliable noninvasive method available for monitoring renal disease in lupus. Twenty-four-hour urine protein levels and urine protein-creatinine ratios appear to correlate well with each other (⬃95% correlation), and represent a well-accepted marker of renal disease that is currently available for clinical use (7). Supplanted with readouts from urinalysis (cells/casts), serum creatinine levels, and kidney biopsy information, the physician is able to plan an appropriate management strategy for the patient. Emerging biomarkers with potential diagnostic value in lupus that have recently been reported include serum levels of various cytokines, mediators, or adhesion molecules (8 –10), gene expression profiles in urine cells (11–13), and urine levels of chemokines (14, 15) and VCAM-1 (16). The present study was initiated based on our recent observation that VCAM-1, P-selectin, soluble TNFR-1 (sTNFR-1), and CXCL16 were elevated in the urine during experimentally induced immune nephritis in mice, correlating well with disease. This experimental model, which is induced by the administration of rabbit anti-glomerular Abs (17), is strain dependent (18, 19). We recently noted that strains that developed more severe disease following the antiglomerular insult exhibited higher urinary levels of VCAM-1, P-selectin, sTNFR-1, and CXCL16 (20). Given that these molecules were previously observed to be generated at least in part in the kidneys and appeared to have pathogenic relevance, it is important to ascertain whether these molecules are also elevated in the urine during spontaneous lupus nephritis. To this end, we monitored the urinary levels of these four molecules in several different strains of lupus mice. Finally, because VCAM-1, P-selectin, sTNFR-1, and CXCL16 appeared to be good markers of renal disease in murine lupus, we proceeded to examine whether these molecules were also elevated in the urine in human lupus nephritis.


7167

Table I. SLE patients and controls examined in this study Subjecta

SLEDAI

Renalb

UrPtn/crc

ANAd

13 14 15 16 17 18 19 21 26 28 29 31 32 43 45 46 48 53 68 69 70 71 72 73 75 89 90 103 104

0 16 2 16 8 8 14 10 8 0 6 5 16 0 0 0 0 18 12 4 2 18 6 14 12 0 8 16 2

0 12 0 0 4 4 8 8 8 0 4 4 12 0 0 0 0 16 12 4 0 12 0 12 8 0 0 0 0

na 217/135 27/249 na 70/64 87/147 na 229/128 385/99

80 320 640 160 160 640 320 320 0 ⬎640 na 0 ⬎640 80 320 160 320 320 320 0 320 0 160 80 na 80 ⬎640 160 80

37/209 36/361 124/106 na 60/123 na 20/275 425/39 179/44 109/61 na 794/251 2/37 63/135 596/115 na 6/45 na na

GN Classe

III, 3/03 V, 6/03 V, 11/04 III, V, 8/04 IV, VI, 2/03 IV, 1/04 IV, 2/03

IV, VI, 1/04 III, 6/04 Vb, 12/02 V, 10/04 III, 11/04 VI, 9/06

ESRD

Subjecta

SLEDAI

Renalb

UrPtn/crc

ANAd

GN Classe

105 109 110 112 113 115 116 120 132 1 2 3 5 6 7 9 11 12 30 33 35 36 39 40 23 24 25 38 41 42

12 6 19 2 2 0 18 0 2 NC NC NC NC NC NC NC NC NC NC NC NC NC NC NC RA RA RA RA RA RA

8 0 12 0 0 0 12 0 0 na na na na na na na na na na na na na na na na na na na na na

223/112 na 121/158 115/260 24/329 26/110 456/198 15/120 29/275 na na na na na na na na na na na na na na na na na na na na na

80 160 na na na 0 640 640 320 na na na na na na na na na na na na na na na 0 0 0 0 0 0

III, 2003

IV, V, 7/02 III, 11/04

a

The first 38 subjects listed represent SLE patients, the next 12 are normal controls (NC), and the last 6 are RA patients. Renal SLEDAI was calculated as detailed in Materials and Methods. Urine protein and creatinine levels shown are in milligrams per deciliter and millimoles per liter, respectively. d Shown are the reciprocals of the dilutions at which the sera were positive for anti-nuclear Ab (ANA). e GN class was assessed according to the World Health Organization scale. The date at which biopsy was done is also shown. na, Not available or not applicable. b c

Materials and Methods Mice C57BL/6 (B6), MRL/lpr, BXSB, and NZM2410 mice were purchased from The Jackson Laboratory or Taconic Farms. B6.Sle1.lpr mice were bred in our mouse colony as described (21). Urine from mice with spontaneous lupus was collected at 6 mo of age, or at the indicated ages. Female mice were used for all experiments, and 24-h urine collections using metabolic cages were conducted for all mice. All animal experiments were conducted in accordance with institutional guidelines.

isotype control Abs, and developed as described previously (21). Briefly, Ag retrieval was performed using a sodium citrate buffer (10 mM sodium citrate, 0.05% Tween 20 (pH 6.0)) in a pressure cooker, protein-blocked for 20 min, and peroxidase-blocked for 20 min. For primary incubation, antiTNFR-1 Ab (0.5 ␮g/ml, 60 min), anti-P-selectin Ab (0.2 ␮g/ml, 60 min), anti-CXCL16 Ab (0.25 ␮g/ml, 60 min), or anti-VCAM-1 Ab (0.3 ␮g/ml, 30 min) were added. This was followed by incubation with biotinylated anti-goat IgG 2 ␮g/ml, ABCelite staining (Vectastain ABCelite kit, HRP), diaminobenzidine chromagen, and hematoxylin. The protein block, diaminobenzidine chromagen, and peroxidase block were purchased from DakoCytomation (catalog no. K4011)).

ELISA detection of sTNFR-1, CXCL16, P-Selectin, and VCAM-1 The following ELISA kits were purchased from R&D Systems and used as dictated by the manufacturer’s protocol (mouse sTNFR-1 DuoSet, catalog no. DY425; mouse CXCL16 DuoSet, catalog no. DY503; mouse P-Selectin DuoSet, catalog no. DY737; mouse VCAM-1 DuoSet, catalog no. DY643; human sTNFR-1 DuoSet, catalog no. DY225; human CXCL16 DuoSet, catalog no. DY1164; human VCAM-1 DuoSet, catalog no. DY809, human P-Selectin immunoassay kit, catalog no. BBE6; human MCP-1, catalog no. DY279; human urokinase-type plasminogen activator (uPAR), catalog no. DY807; human IL-1, catalog no. DY201; human IL-6, catalog no. DR206). All urine samples were diluted 1/5 or higher for the ELISA, and the concentrations of the respective molecules were ascertained using manufacturer supplied standards. Serum samples were diluted 1/100 for the ELISA. Likewise, all culture supernatants were diluted 1/2 or 1/10 before assaying for the released mediators. The absolute concentrations were determined by multiplying the ELISA-determined values by the respective dilutions factors.

Immunohistochemistry Kidney sections obtained from 2- or 6-mo-old mice were stained with the following primary Abs: goat anti-mouse CXCL16 Ab (no. AF503; R&D Systems), goat anti-mouse TNFR-1 Ab (no. BAF425; R&D Systems), goat anti-mouse VCAM-1 (no. BAF643; R&D Systems) Ab, or goat anti-mouse P-selectin Ab (no. SC-6941; Santa Cruz Biotechnology) or the respective

Patient samples Urine and sera from healthy adults (n ⫽ 15, average age ⫽ 43), SLE patients (n ⫽ 38; with or without active nephritis; average age ⫽ 34), and rheumatoid arthritis (RA) patients (n ⫽ 6, average age ⫽ 61) were drawn at the Albert Einstein College of Medicine, in accordance with institutional review board approved guidelines. All patients and controls (except for one) were females. Most of the individuals in all groups were of Hispanic or African-American origin. Disease activity was gauged using SLE disease activity index (SLEDAI) scores (22). The patients exhibited a wide spread of SLEDAI scores displayed in Table I. The degree of lupus nephritis was determined by standard clinical (serology, renal function, urinary sediment, and protein) and/or pathological (kidney biopsy) criteria. All patient studies were conducted with informed consent and institutionally approved guidelines.

Renal SLEDAI Kidney disease activity was assessed using the renal SLEDAI score that consists of the four kidney-related criteria of the SLEDAI (22) (i.e., hematuria, pyuria, proteinuria, and urinary casts). The presence of each one of these four parameters yields a score of 1– 4 points, depending on the severity; thus, the renal SLEDAI score can range from 0 (in nonactive renal disease) to a maximal score of 16.

7168

VCAM-1, P-SELECTIN, sTNFR-1, AND CXCL16 IN LUPUS URINE

FIGURE 1. VCAM-1, P-selectin, sTNFR-1, and CXCL16 are elevated in the urine in spontaneous lupus nephritis in three different mouse models of lupus. The levels of the above molecules were determined by ELISA in 24-h urine samples obtained from 6-mo-old B6.Sle1.lpr, MRL.lpr, and NZM2410 lupus mice (all of which exhibited proteinuria and glomerulonephritis), as well as 6-mo-old B6 controls (A). Indicated p values pertain to Student’s t test or Mann-Whitney U test comparisons of B6 urine vs the other groups (ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01; ⴱⴱⴱ, p ⬍ 0.001). n ⫽ 5 mice/group. For each urine sample depicted in Fig. 1A, the 24-h urine level of total protein was plotted against the 24-h urine levels of the four molecules (B). Typically, healthy B6 mice do not excrete ⬎1 mg of protein over a 24-h period (18, 19).

Statistics Groups were compared against each other using the Student t test where the data were normally distributed. Otherwise, the nonparametric Mann-Whitney U test was used. Most statistical tests including correlation coefficients (R) were calculated using Excel or GraphPad Prism. The nonparametric receiver operating characteristic (ROC) curves were calculated using Intercooled Stata version 9.2 (StataCorp). The area under the ROC curves (“AUC”) were reflective of the specificity and sensitivity of the marker, with values of 0.7– 0.8, 0.8 – 0.9, and 0.9 –1.0 representing “acceptable,” “excellent,” and “outstanding” discriminatory potential, respectively.

Results VCAM-1, P-selectin, TNFR-1, and CXCL16 are elevated in the urine in spontaneous lupus nephritis B6.Sle1.lpr, MRL.lpr, and NZM2410 mice develop high levels of anti-dsDNA and anti-glomerular autoantibodies, accompanied by severe lupus nephritis, and early mortality (21, 23–25). When the above quartet of molecules was assayed, it was evident that urine from these strains of mice suffering from spontaneous lupus nephritis also harbored increased levels of these molecules (Fig. 1A). Indeed, the 24-h urine levels of these molecules were reflective of the extent of renal disease in these mice (Fig. 1B). Specifically, the urine levels of protein correlated well with the corresponding urine levels of VCAM-1 (R ⫽ 0.8, p ⬍ 0.001), CXCL16 (R ⫽ 0.78, p ⬍

0.001), and P-selectin (R ⫽ 0.75, p ⬍ 0.001), but less so with TNFR-1 levels (R ⫽ 0.57, p ⬍ 0.001) (Fig. 1B). Urinary VCAM-1, P-selectin, sTNFR-1, and CXCL16 are higher than the corresponding serum levels Given the observation that VCAM-1, P-selectin, sTNFR-1, and CXCL16 were elevated in the urine of lupus mice, we next examined the degree to which these molecules may be serum-derived, focusing initially on the MRL.lpr strain. It was clear that with disease, the serum levels of all four molecules increased ⬃2- to 4-fold, relative to predisease 2-mo-old MRL.lpr mice and B6 controls (Fig. 2). It is remarkable that the serum levels of VCAM-1 and P-selectin in 6-moold MRL.lpr mice ranged from 1 to 5 mg/ml and 0.25– 0.75 mg/ml, respectively, levels that were significantly higher than the levels of the other two molecules. These findings raise the possibilities that the urinary content of these molecules may in part be serum derived. To ascertain whether any of these molecules were further enriched in the urine relative to the serum, the urine-serum ratios of these molecules were next calculated, after normalizing both values against albumin. The degree of enrichment of the four molecules in urine followed this order: VCAM-1 ⬎ P-selectin ⬎ CXCL16 ⬎ TNFR-1 (Fig. 2). Whereas VCAM-1 was enriched 2to 10-fold in the urine relative to the serum in 6-mo-old MRL.lpr mice, TNFR-1 was hardly so. P-selectin and CXCL16 exhibited


7169

FIGURE 2. Urinary VCAM-1, P-selectin, sTNFR-1, and CXCL16 are higher than the corresponding serum levels. Serum and urine levels of the above four molecules were determined by ELISA, using samples from 2-mo-old B6, 2-mo-old MRL.lpr, and 6-mo-old MRL.lpr mice. Indicated p values pertain to Student’s t test or Mann-Whitney U test comparisons of B6 urine vs the other groups (indicated below the x-axis), or young vs old MRL.lpr samples, as indicated above each plot (ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01; ⴱⴱⴱ, p ⬍ 0.001). n ⫽ 12 mice/group.

intermediate degrees of enrichment in the urine, suggesting that they maybe partly serum derived and partly kidney derived. Similar urinary enrichment of these molecules were also noted in two additional strains which develop spontaneous lupus, B6.Sle1.lpr and BXSB, in an age- and disease-dependent fashion, as illustrated for VCAM-1 and CXCL16 (Fig. 3). VCAM-1, TNFR-1, P-selectin, and CXCL16 are hyperexpressed within the diseased kidneys in lupus Given that most of the above molecules were enriched in the urine during the course of lupus, we wondered whether they may be originating from the kidneys. To ascertain this, we next examined the levels of these molecules in renal cortical lysates from MRL.lpr and BXSB mice with lupus nephritis. As depicted in Fig. 4, all four molecules were expressed in the renal cortical lysates from both these strains with spontaneous lupus nephritis. Immunohistochemistry studies further highlighted the fine localization of these four molecules within the renal cortex (Fig. 5). Diseased MRL.lpr kidneys exhibited VCAM-1 staining in the peritubular capillaries, vascular bundles, endothelium, and scattered glomerular cells as well as in interstitial inflammatory cells. Staining of the same cells at a weaker level was noted in 2-mo-old MRL.lpr kidneys before disease onset (Fig. 5 and data not shown). The remaining three molecules, P-selectin, sTNFR-1, and CXCL16, showed a fairly similar distribution, being observed in the distal tubules, collecting ducts, endothelium, the interstitial inflammatory cells, and scattered glomerular cells. Once again, staining was more prominent in

diseased MRL.lpr kidneys, compared with 2-mo-old MRL.lpr or B6 control kidneys (Fig. 5). VCAM-1, TNFR-1, P-selectin, and CXCL16 are also elevated in the urine of SLE patients To determine whether this quartet of molecules was elevated in the urine of SLE patients, urine samples from 38 SLE patients with a spread of SLEDAI scores (Table I), 6 RA patients, and 15 normal healthy controls were studied for the expression of these four molecules, using commercially available ELISA kits. As shown in Fig. 6A, urine from patients with SLE showed significantly higher levels of urinary VCAM-1, CXCL16, TNFR-1, and P-selectin, compared with urine from RA patients and normal controls. Importantly, urine levels of these molecules showed excellent correlation with concurrent urinary protein-creatinine ratios (Fig. 6B), and SLEDAI scores (Fig. 6C). Not surprisingly, the urinary levels of all four molecules exhibited excellent ROC profiles, with AUC values (ranging from 0.76 – 0.89) reflecting high degrees of sensitivity and specificity for distinguishing SLE patients from healthy controls (Fig. 7). Because patients with elevated SLEDAI may or may not necessarily have renal disease, we next focused on SLE patients with active nephritis. Importantly, SLE patients with active nephritis exhibited the highest levels of these molecules in the urine, particularly VCAM-1 and CXCL16. Thus, patients with increased SLEDAI and active renal disease (SLEDAI and renal SLEDAI ⬎0; n ⫽ 18) had significantly higher levels of urinary VCAM-1/ creatinine (618 vs 370 pg/mM, p ⬍ 0.02) and CXCL16/creatinine

7170


FIGURE 3. The urinary enrichment of VCAM-1 and CXCL16 is also noted in other strains with spontaneous lupus nephritis. Serum and urine levels of VCAM-1 and CXCL16 were determined by ELISA, using samples from 2- to 5-mo-old B6, B6.Sle1.lpr, and BXSB mice, as indicated. Displayed p values pertain to the Student t test or Mann-Whitney U test comparisons of young vs old mice (ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01; ⴱⴱⴱ, p ⬍ 0.001); n ⫽ 6 mice/group.

(154 vs 85 pg/mM, p ⬍ 0.04), compared with patients with elevated SLEDAI but without active nephritis (SLEDAI ⱖ6, renal SLEDAI ⫽ 0; n ⫽ 5; Fig. 8). Because both these groups of patients (with nephritis vs those without) had SLEDAI scores that were not

FIGURE 4. VCAM-1, TNFR-1, P-selectin, and CXCL16 are hyperexpressed within the diseased kidneys in lupus. Renal cortical lysates were prepared from B6 control, as well as BXSB and MRL.lpr mice, aged 5– 6 mo. The levels of the four molecules were determined by ELISA. Indicated p values pertain to Student’s t test or Mann-Whitney U test comparisons of B6 vs other strains (ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01; ⴱⴱⴱ, p ⬍ 0.001). n ⫽ 6 mice/group. Group means are indicated by the horizontal bars.

FIGURE 5. CXCL16, p-selectin, sTNFR-1, and VCAM-1 may be expressed in the kidneys. Immunostaining of mouse kidneys from 2- and 7-mo-old MRL.lpr mice was performed using Abs against the above four molecules. Shown figures are representative of the histology seen in five mice per group.

statistically different (average of 12.5 and 10.4, respectively), the differences in the urinary levels of these molecules were not simply due to differences in SLEDAI scores. Examined using an alternative approach, urinary VCAM-1, sTNFR-1, and CXCL16 exhibited “outstanding” sensitivity and specificity AUC values (0.91– 0.93) for discriminating SLE patients with active renal disease, as is evident from the ROC curves of these molecules (Fig. 9 and data not plotted). We next examined the levels of VCAM-1, CXCL16, TNFR-1, and P-selectin in the sera of these patients after distinguishing those with active disease from the rest. Several important observations were made. First, the levels of all molecules were elevated in the sera of SLE patients with active but not inactive disease (Fig. 10). Second, it was clear that VCAM-1, CXCL16, TNFR-1, and P-selectin were further enriched in the urine, relative to the serum, only among the lupus patients with active disease (Fig. 10), closely mirroring the findings in lupus mice. Finally, we compared the performance of the above markers to that of four additional molecules, IL-1, IL-6, uPAR, and MCP-1. The first three represent inflammatory mediators that have previously been studied in several end-organ diseases, whereas MCP-1 is a chemokine that has previously been documented to be a potential biomarker in lupus nephritis (15). IL-1 and IL-6 were barely detected in the urine samples studied (data not shown); in contrast, significant levels of uPAR were detected in these samples. However, as shown in Fig. 11A, urinary uPAR levels were not significantly different in the different SLE patient groups (e.g., nephritis vs non-nephritis, or active vs


7171

FIGURE 6. Urine VCAM-1, P-selectin, sTNFR-1, and CXCL16 are also elevated in human lupus nephritis. A, Spot urine samples from 38 SLE patients (Table I), 15 healthy adults, and 6 RA patients were assayed for the levels of VCAM-1, P-selectin, sTNFR-1, and CXCL16. Indicated p values pertain to Student’s t test or Mann-Whitney U test comparisons of the urinary levels of each molecule in patients against that in normal controls (indicated below the x-axis), or SLE vs RA, as indicated at the top (ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01; ⴱⴱⴱ, p ⬍ 0.001). Shown also are the correlation profiles of these urinary molecules (normalized to urine creatinine) against the corresponding urine protein-creatinine ratios (B) and SLEDAI scores (C). R: correlation coefficient.

inactive disease), and were barely different from the corresponding levels in healthy controls. In contrast, urinary MCP-1 was higher in the urine of SLE patients compared with normal urine (Fig. 11B). However, these levels were not significantly higher in the urine of nephritic SLE patients

FIGURE 7. Specificity/sensitivity ROC curves pertaining to the use of urinary VCAM-1 (A), CXCL16 (B), sTNFR-1 (C), and P-selectin (D) for discriminating SLE patients from normal controls. The data shown in Fig. 6 were used to generate the shown ROC curves, which plot the specificity and sensitivity profiles of the indicated urinary molecules in distinguishing SLE patients from normal controls. Indicated are the corresponding AUC values.

(Fig. 11B) compared with the urinary levels in inactive or nonrenal SLE. The corresponding ROC AUC values for urinary MCP-1 and uPAR for distinguishing active renal lupus from nonrenal lupus were 0.78 and 0.77, respectively (data not plotted).

7172


FIGURE 8. SLE patients with active nephritis exhibit the highest levels of urine VCAM-1, P-selectin, sTNFR-1, and CXCL16. Urine levels of the indicated molecules were assayed in SLE patients and normal controls. The SLE patients were divided into four groups: (1) inactive SLE (SLEDAI ⫽ 0); (2) mild-active, nonrenal SLE (SLEDAI ⫽ 1–5, renal SLEDAI ⫽ 0); (3) severe-active, nonrenal SLE (SLEDAI ⱖ6; renal SLEDAI ⫽ 0); (4) active renal SLE (renal SLEDAI ⬎0). Indicated p values pertain to Student’s t test or Mann-Whitney U test comparisons of each subject group vs “mild-active, nonrenal SLE” (ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01; ⴱⴱⴱ, p ⬍ 0.001).

Discussion Given that early intervention in lupus nephritis is associated with better outcome, it is imperative that we diagnose the disease activity in the kidney as early as possible. Currently, urinary albumin levels, supplanted by additional laboratory tests, constitute the cornerstone of diagnostic screening in lupus nephritis. There have been isolated reports that the urinary levels of certain chemokines, notably MCP-1, or adhesion molecules, notably VCAM-1, might have the potential to serve as additional biomarkers for lupus nephritis (8, 14 –16). The present study supplements this list with additional excreted markers, P-selectin, sTNFR-1, and CXCL16, which are reproducibly excreted in the urine in spontaneous murine and human lupus nephritis. Given that these molecules may also be expressed within the diseased kidneys, these molecules may also be playing critical roles in disease pathogenesis. VCAM-1 is an adhesion molecule that is expressed on a large number of cell types including macrophages, dendritic cells (DCs), endothelial cells, etc., and plays an important role in cell recruitment into tissues mediated by VCAM-1:VLA4 interactions. Increased expression of VCAM-1 has also been noted in arthritis and other immune disorders where it has also been identified as a therapeutic target (26, 27). It has previously been observed to be expressed in the kidneys, both in murine and human lupus (28 –30), and this is reinforced by our present study that shows VCAM-1 staining in murine lupus kidneys. Our findings resonate well with a previous report of increased urinary VCAM-1 in lupus nephritis

FIGURE 9. Specificity/sensitivity ROC curves pertaining to the use of urinary VCAM-1 (A) and CXCL16 (B) for discriminating SLE patients with active renal disease. The data shown in Fig. 8 were used to generate the shown ROC curves, which plot the specificity and sensitivity profiles of the indicated urinary molecules in distinguishing SLE patients with active renal disease from all other SLE patients. Indicated are the corresponding AUC values.

(16). Though serum levels of VCAM-1 increased significantly with disease both in murine and human lupus, the kidney appeared to be an additional source of VCAM-1, based on the immunohistochemistry and renal expression studies and the enrichment in the urine particularly in active disease. The present findings raise hope that VCAM-1 may be a reliable urinary marker of active disease both in mice and patients with nephritis, thus confirming earlier reports. P-selectin is an adhesion molecule that is expressed on platelets and other cell types, and has been reported to be elevated in other autoinflammatory diseases, where it has also been recognized as a circulating disease marker (31–33). It has been observed to be expressed in lupus kidneys, both in mice and in patients (34, 35), but it has not been specifically measured in lupus urine. Hence, this represents the first report of elevated urinary P-selectin in lupus nephritis. Though serum levels of P-selectin increased significantly with disease both in murine and human lupus, the kidney appeared to be an additional source of P-selectin based on the immunohistochemistry and renal expression studies and the enrichment in the urine, in the murine studies. TNFR-1 is a cytokine receptor that plays various functions in the immune system. Importantly, increased serum TNFR-1 has been noted to be increased in human SLE (36, 37). There is a single report of its being hyperexpressed in the kidneys in proliferative lupus nephritis (38), but it has not been examined in the


7173

FIGURE 10. Urine vs serum levels of VCAM-1, P-selectin, sTNFR-1, and CXCL16 in human lupus. Serum levels of the above four molecules were also assayed in SLE patients with inactive SLE (SLEDAI ⫽ 0; n ⫽ 5, randomly selected from the patients listed in Table I) and in patients with “active SLE” (SLEDAI ⱖ12; n ⫽ 12, randomly selected from the patients listed in Table I). Displayed are the absolute levels (left column) as well as the albuminnormalized values (middle column) for the four molecules assayed. Displayed in the right column are the urine-serum ratios of these four molecules, after normalization against the corresponding albumin levels. Indicated p values pertain to Student’s t test or MannWhitney U test comparisons of patients vs “normal” (indicated below the x-axis), or “inactive” vs “active” patients, as indicated above each plot (ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01; ⴱⴱⴱ, p ⬍ 0.001).

urine as yet. Hence, this represents the first report of elevated urinary sTNFR-1 in lupus nephritis. However, urinary TNFR-1 in murine and human lupus nephritis may be originating from the serum at least in part, based on our findings, consistent with our findings in the experimental anti-glomerular basement membrane (GBM) model (20). CXCL16 is a chemokine, reported to be expressed on DCs and macrophages, that is important for the recruitment of T cells and NK T cells into various tissues, and for cell:cell interactions via the CXCR6 counterreceptor. It also appears to be important in arthritis (39 – 41); however, it has not been studied before in lupus, either in the kidneys or in the urine. Hence, this represents the first report of its increased presence in lupus urine. Though serum levels of CXCL16 increased significantly with disease both in murine and human lupus, urinary CXCL16 may be at least partly generated in the kidneys based on the renal expression studies. The immunohistochemistry studies revealed that the above quartet of molecules are expressed by multiple cell types in the kidneys, including endothelial cells, tubular cells, glomerular cells as well as infiltrating leukocytes. We believe the inflamed kidney represents an important source of these molecules in the urine, based on the following observations. First, these molecules are enriched in the urine, relative to the blood, both in murine and human lupus (Figs. 3 and 10). Second, these molecules are expressed within the renal parenchyma, based on findings from two orthogonal studies (Figs. 4 and 5). Third, patients with active renal lupus had the highest urinary levels of these molecules (Figs. 6 and 8). Finally, these molecules have high specificity and sensitivity in discriminating active renal

lupus from nonrenal disease, as indicated by their high ROC AUC values (Fig. 9). In addition, it appears likely that the above molecules may be serum derived in part, arising as a consequence of systemic inflammation. To begin with, these molecules are elevated in the serum, both in mice and patients with SLE. Additionally, RA patients exhibit urinary (and serum) levels of these molecules that are significantly higher than the levels in normal mice (Fig. 6). These observations suggest that systemic inflammation may result in increased levels of circulating VCAM-1, P-selectin, sTNFR-1, and CXCL16. Given that levels of these molecules correlate well with proteinuria or disease severity both in mice and patients, it is imperative that we assess the diagnostic use of assaying these molecules in the urine. Even if the serum may constitute a significant source of these molecules in the urine, the ability to detect these molecules in the urine could have a tremendous impact on clinical diagnostics. Not only is urine a far more convenient body fluid to procure, in some clinical settings it may be the only fluid available (e.g., in field hospitals or from pediatric patients). However, it is clear that serum is not the only source of urinary VCAM-1, P-selectin, CXCL16, and sTNFR-1 in lupus. The observation that these molecules are enriched in the urine (relative to the serum, after normalizing against albumin) and the demonstration of the presence of these molecules in the renal parenchyma using two orthogonal approaches suggest that these molecules may also be generated in situ in the diseased kidneys. Given the diagnostic potential of these molecules, it is imperative to systematically monitor the urinary levels of these molecules

7174

VCAM-1, P-SELECTIN, sTNFR-1, AND CXCL16 IN LUPUS URINE (26, 27). Likewise, it has been suggested that glomerular disease may be amenable for therapy by blocking P-selectin (42). Based on our Ab blocking studies in the experimental antiGBM nephritis model, CXCL16 may turn out to be yet another therapeutic target in nephritis (20). Finally, it remains to be seen whether the coordinate blocking of one or more of these molecules may offer therapeutic relief that is even more superior in lupus nephritis.

Acknowledgments We acknowledge Madhavi Bhaskharabhatla and Amanda Leon for their expert technical assistance.

Disclosures The authors have no financial conflict of interest.

References

FIGURE 11. Urine uPAR (A) and MCP-1 (B) levels in SLE patients with or without active nephritis. Urine levels of the indicated molecules were assayed in SLE patients and normal controls. The SLE patients were divided into four groups: 1) inactive SLE (SLEDAI ⫽ 0); 2) mildactive, nonrenal SLE (SLEDAI ⫽ 1–5, renal SLEDAI ⫽ 0); 3) severe-active, nonrenal SLE (SLEDAI ⱖ6; renal SLEDAI ⫽ 0);) active renal SLE (renal SLEDAI ⬎0). Indicated p values pertain to Student’s t test or Mann-Whitney U test comparisons of each subject group vs “mild-active, nonrenal SLE.” NS, Not significant.

longitudinally, with supporting biopsy information. Recent longitudinal studies in human lupus have implicated urinary MCP-1 as a potential early biomarker of lupus nephritis (15). Our studies comparing the performance of urinary VCAM-1, sTNFR-1, and CXCL16 to that of urinary MCP-1 in the same cohort of SLE patients indicate that the former molecules are truly promising candidates for systematic evaluation as potential biomarkers in lupus nephritis. Given the outcomes of these comparative cross-sectional studies, the novel urinary molecules described in this study are anticipated to perform at least as well as MCP-1 in longitudinal studies. The molecules studied in this communication may also have important roles in disease pathogenesis, based on several counts. First, limited information is already available from the published literature concerning the pathogenic relevance of some of these molecules in other disease settings, as partly discussed above. Second, these molecules are enriched in the urine, particularly in mice and patients with severe lupus nephritis, an observation that begs the question of whether these molecules are actually mediating the pathogenesis of nephritis. Third, we have recently demonstrated the pathogenic role of CXCL16 in experimental anti-GBM disease (20). Given these promising leads, more work needs to be done to elucidate the cellular origins of these molecules within the diseased kidneys, and to establish the degree to which these different molecules are required for renal disease to ensue in lupus. Finally, given the observation by ourselves and others (20) that at least some of these molecules may play critical roles in the pathogenesis of nephritis, they also represent potential targets for therapeutic intervention. Indeed, VCAM-1 has been suggested as a potential therapeutic target in inflammatory end-organ disease

1. Hahn, B. H. 1998. Antibodies to DNA. N. Engl. J. Med. 338: 1359 –1368. 2. Kotzin, B. L. 1996. Systemic lupus erythematosus. Cell 85: 303–306. 3. Houssiau, F. A. 2004. Management of lupus nephritis: an update. J. Am. Soc. Nephrol. 15: 2694 –2704. 4. Fiehn, C., Y. Hajjar, K. Mueller, R. Waldherr, A. D. Ho, and K. Andrassy. 2003. Improved clinical outcome of lupus nephritis during the past decade: importance of early diagnosis and treatment. Ann. Rheum. Dis. 62: 435– 439. 5. Moroni, G., S. Pasquali, S. Quaglini, G. Banfi, S. Casanova, M. Maccario, P. Zucchelli, and C. Ponticelli. 1999. Clinical and prognostic value of serial renal biopsies in lupus nephritis. Am. J. Kidney Dis. 34: 530 –539. 6. Illei, G. G., and P. E. Lipsky. 2004. Biomarkers in systemic lupus erythematosus. Curr. Rheumatol. Rep. 6: 382–390. 7. Christopher-Stine, L., M. Petri, B. C. Astor, and D. Fine. 2004. Urine proteinto-creatinine ratio is a reliable measure of proteinuria in lupus nephritis. J. Rheumatol. 31: 1557–1559. 8. Spronk, P. E., H. Bootsma, M. G. Huitema, P. C. Limburg, and C. G. Kallenberg. 1994. Levels of soluble VCAM-1, soluble ICAM-1, and soluble E-selectin during disease exacerbations in patients with systemic lupus erythematosus (SLE); a long term prospective study. Clin. Exp. Immunol. 97: 439 – 444. 9. Belmont, H. M., J. Buyon, R. Giorno, and S. Abramson. 1994. Up-regulation of endothelial cell adhesion molecules characterizes disease activity in systemic lupus erythematosus: the Shwartzman phenomenon revisited. Arthritis Rheum. 37: 376 –383. 10. Manzi, S., J. S. Navratil, M. J. Ruffing, C. C. Liu, N. Danchenko, S. E. Nilson, S. Krishnaswami, D. E. King, A. H. Kao, and J. M. Ahearn. 2004. Measurement of erythrocyte C4d and complement receptor 1 in systemic lupus erythematosus. Arthritis Rheum. 50: 3596 –3604. 11. Nakamura, T., C. Ushiyama, S. Suzuki, M. Hara, N. Shimada, K. Sekizuka, I. Ebihara, and H. Koide. 2000. Urinary podocytes for the assessment of disease activity in lupus nephritis. Am. J. Med. Sci. 320: 112–116. 12. Chan, R. W., F. M. Lai, E. K. Li, L. S. Tam, T. Y. Wong, C. Y. Szeto, P. K. Li, and C. C. Szeto. 2004. Expression of chemokine and fibrosing factor messenger RNA in the urinary sediment of patients with lupus nephritis. Arthritis Rheum. 50: 2882–2890. 13. Chan, R. W., L. S. Tam, E. K. Li, F. M. Lai, K. M. Chow, K. B. Lai, P. K. Li, and C. C. Szeto. 2003. Inflammatory cytokine gene expression in the urinary sediment of patients with lupus nephritis. Arthritis Rheum. 48: 1326 –1331. 14. Tucci, M., E. V. Barnes, E. S. Sobel, B. P. Croker, M. S. Segal, W. H. Reeves, and H. B. Richards. 2004. Strong association of a functional polymorphism in the monocyte chemoattractant protein 1 promoter gene with lupus nephritis. Arthritis Rheum. 50: 1842–1849. 15. Rovin, B. H., H. Song, D. J. Birmingham, L. A. Hebert, C. Y. Yu, and H. N. Nagaraja. 2005. Urine chemokines as biomarkers of human systemic lupus erythematosus activity. J. Am. Soc. Nephrol. 16: 467– 473. 16. Molad, Y., E. Miroshnik, J. Sulkes, S. Pitlik, A. Weinberger, and Y. Monselise. 2002. Urinary soluble VCAM-1 in systemic lupus erythematosus: a clinical marker for monitoring disease activity and damage. Clin. Exp. Rheumatol. 20: 403– 406. 17. Masugi, M. 1934. Glomerulonephritis durch das spezifische antinierenserumein Beitrag zur Pathogenese der diffusen glomerulonephritis. Beitr. Pathol. Anat. Allgemein. Pathol. 92: 429. 18. Xie, C., X. J. Zhou, X. Liu, and C. Mohan. 2003. Enhanced susceptibility to end-organ disease in the lupus-facilitating NZW mouse strain. Arthritis Rheum. 48: 1080 –1092. 19. Xie, C., R. Sharma, H. Wang, X. J. Zhou, and C. Mohan. 2004. Strain distribution pattern of susceptibility to immune-mediated nephritis. J. Immunol. 172: 5047–5055. 20. Wu, T., X. Chun, M. Bhaskharabatla, A. Leon, X. J. Zhou, C. Putterman, and C. Mohan. 2007. Excreted urinary mediators in immune nephritis with potential pathogenic significance. Arthritis Rheum. 56: 949 –959. 21. Shi, X., C. Xie, D. Kreska, J. A. Richardson, and C. Mohan. 2002. Genetic dissection of SLE: SLE1 and FAS impact alternate pathways leading to lymphoproliferative autoimmunity. J. Exp. Med. 196: 281–292. 22. Gladman, D. D., D. Ibanez, and M. B. Urowitz. 2002. Systemic lupus erythematosus disease activity index 2000. J. Rheumatol. 29: 288 –291.

The Journal of Immunology 23. Rudofsky, U. H., B. D. Evans, S. L. Balaban, V. D. Mottironi, and A. E. Gabrielsen. 1993. Differences in expression of lupus nephritis in New Zealand mixed H-2z homozygous inbred strains of mice derived from New Zealand black and New Zealand white mice: origins and initial characterization. Lab. Invest. 68: 419 – 426. 24. Cohen, P. L., and R. A. Eisenberg. 1991. lpr and gld: single gene models of systemic autoimmunity and lymphoproliferative disease. Annu. Rev. Immunol. 9: 243–269. 25. Fu, Y., C. Xie, M. Yan, Q. Li, J. W. Joh, C. Lu, and C. Mohan. 2005. The lipopolysaccharide-triggered mesangial transcriptome: evaluating the role of interferon regulatory factor-1. Kidney Int. 67: 1350 –1361. 26. Matsuyama, T., and A. Kitani. 1996. The role of VCAM-1 molecule in the pathogenesis of rheumatoid synovitis. Hum. Cell. 9: 187–192. 27. Foster, C. A. 1996. VCAM-1/␣ 4-integrin adhesion pathway: therapeutic target for allergic inflammatory disorders. J. Allergy Clin. Immunol. 98: S270 –S277. 28. Nakatani, K., H. Fujii, H. Hasegawa, M. Terada, N. Arita, M. R. Ito, M. Ono, S. Takahashi, K. Saiga, S. Yoshimoto, et al. 2004. Endothelial adhesion molecules in glomerular lesions: association with their severity and diversity in lupus models. Kidney Int. 65: 1290 –1300. 29. Seron, D., J. S. Cameron, and D. O. Haskard. 1991. Expression of VCAM-1 in the normal and diseased kidney. Nephrol. Dial. Transplant. 6: 917–922. 30. Wuthrich, R. P. 1992. Vascular cell adhesion molecule-1 (VCAM-1) expression in murine lupus nephritis. Kidney Int. 42: 903–914. 31. Kappelmayer, J., B. Nagy, Jr., K. Miszti-Blasius, Z. Hevessy, and H. Setiadi. 2004. The emerging value of P-selectin as a disease marker. Clin. Chem. Lab. Med. 42: 475– 486. 32. Geng, J. G., M. Chen, and K. C. Chou. 2004. P-selectin cell adhesion molecule in inflammation, thrombosis, cancer growth and metastasis. Curr. Med. Chem. 11: 2153–2160. 33. Sfikakis, P. P., D. Charalambopoulos, G. Vaiopoulos, and M. Mavrikakis. 1999. Circulating P- and L-selectin and T-lymphocyte activation and patients with autoimmune rheumatic diseases. Clin. Rheumatol. 18: 28 –32. 34. Segawa, C., T. Wada, M. Takaeda, K. Furuichi, I. Matsuda, Y. Hisada, S. Ohta, K. Takasawa, S. Takeda, K. Kobayashi, and H. Yokoyama. 1997. In situ expres-

7175

35.

36.

37.

38.

39.

40.

41.

42.

sion and soluble form of P-selectin in human glomerulonephritis. Kidney Int. 52: 1054 –1063. Takaeda, M., H. Yokoyama, C. Segawa-Takaeda, T. Wada, and K. Kobayashi. 2002. High endothelial venule-like vessels in the interstitial lesions of human glomerulonephritis. Am. J. Nephrol. 22: 48 –57. Davas, E. M., A. Tsirogianni, I. Kappou, D. Karamitsos, I. Economidou, and P. C. Dantis. 1999. Serum IL-6: TNF␣, p55 srTNF␣, p75srTNF␣, srIL-2␣ levels and disease activity in systemic lupus erythematosus. Clin. Rheumatol. 18: 17–22. Gabay, C., N. Cakir, F. Moral, P. Roux-Lombard, O. Meyer, J. M. Dayer, T. Vischer, H. Yazici, and P. A. Guerne. 1997. Circulating levels of tumor necrosis factor soluble receptors in systemic lupus erythematosus are significantly higher than in other rheumatic diseases and correlate with disease activity. J. Rheumatol. 24: 303–308. Aten, J., A. Roos, N. Claessen, E. J. Schilder-Tol, I. J. Ten Berge, and J. J. Weening. 2000. Strong and selective glomerular localization of CD134 ligand and TNF receptor-1 in proliferative lupus nephritis. J. Am. Soc. Nephrol. 11: 1426 –1438. Chandrasekar, B., S. Bysani, and S. Mummidi. 2004. CXCL16 signals via Gi, phosphatidylinositol 3-kinase: Akt, I␬B kinase, and nuclear factor-␬B and induces cell-cell adhesion and aortic smooth muscle cell proliferation. J. Biol. Chem. 279: 3188 –3196. Shimaoka, T., T. Nakayama, N. Fukumoto, N. Kume, S. Takahashi, J. Yamaguchi, M. Minami, K. Hayashida, T. Kita, J. Ohsumi, et al. 2004. Cell surface-anchored SR-PSOX/CXC chemokine ligand 16 mediates firm adhesion of CXC chemokine receptor 6-expressing cells. J. Leukocyte Biol. 75: 267–274. Gough, P. J., K. J. Garton, P. T. Wille, M. Rychlewski, P. J. Dempsey, and E. W. Raines. 2004. A disintegrin and metalloproteinase 10-mediated cleavage and shedding regulates the cell surface expression of CXC chemokine ligand 16. J. Immunol. 172: 3678 –3685. Matsuda, M., K. Shikata, F. Shimizu, Y. Suzuki, M. Miyasaka, H. Kawachi, H. Kawashima, J. Wada, H. Sugimoto, Y. Shikata, et al. 2002. Therapeutic effect of sulphated hyaluronic acid, a potential selectin-blocking agent, on experimental progressive mesangial proliferative glomerulonephritis. J. Pathol. 198: 407– 414.