Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Kidney Int. Author manuscript; available in PMC 2009 April 10.
Published in final edited form as:
PMCID: PMC2667626

Prediction of urinary protein markers in lupus nephritis



Lupus nephritis is divided into six classes and scored according to activity and chronicity indices based on histologic findings. Treatment differs based on the pathologic findings. Renal biopsy is currently the only way to accurately predict class and activity and chronicity indices. We propose to use patterns of abundance of urine proteins to identify class and disease indices.


Urine was collected from 20 consecutive patients immediately prior to biopsy for evaluation of lupus nephritis. The International Society of Nephrology/Renal Pathology Society (ISN/RPS) class of lupus nephritis, activity, and chronicity indices were determined by a renal pathologist. Proteins were separated by two-dimensional gel electrophoresis. Artificial neural networks were trained on normalized spot abundance values.


Biopsy specimens were classified in the database according to ISN/RPS class, activity, and chronicity. Nine samples had characteristics of more than one class present. Receiver operating characteristic (ROC) curves of the trained networks demonstrated areas under the curve ranging from 0.85 to 0.95. The sensitivity and specificity for the ISN/RPS classes were class II 100%, 100%; III 86%, 100%; IV 100%, 92%; and V 92%, 50%. Activity and chronicity indices had r values of 0.77 and 0.87, respectively. A list of spots was obtained that provided diagnostic sensitivity to the analysis.


We have identified a list of protein spots that can be used to develop a clinical assay to predict ISN/RPS class and chronicity for patients with lupus nephritis. An assay based on antibodies against these spots could eliminate the need for renal biopsy, allow frequent evaluation of disease status, and begin specific therapy for patients with lupus nephritis.

Keywords: lupus nephritis, biomarkers, urine, electrophoresis, two-dimensional gel

Patients with systemic lupus erythematosus (SLE) frequently develop renal disease that may lead to end-stage renal disease (ESRD). In 2002, 1100 patients developed ESRD secondary to SLE, a 12% increase from the previous year [1]. The renal disease of SLE has been divided into six classes by the International Society of Nephrology/Renal Pathology Society (ISN/RPS): class I, mesangial immune deposits; class II, mesangial immune deposits with mesangial hypercellularity; class III, proliferative glomerulonephritis involving <50% of glomeruli; class IV, proliferative glomerulonephritis involving ≥50% of glomeruli; class V, membranous nephritis; and class VI, advanced sclerosing lesions [2]. More than one type of renal disease can be present simultaneously, and the class of disease can change over time [3]. The histologic disease can also be classified according to activity and chronicity indices [4]. Activity and chronicity indices quantitate damage that is thought to be recent and potentially reversible or longstanding and irreversible. The prognosis and treatment of lupus nephritis are different according to the ISN/RPS class of disease present, activity, and chronicity [58]. Identification of ISN/RPS class and activity and chronicity indices requires renal biopsy.

A diagnostic test based on urine proteins that could identify the ISN/RPS class of lupus nephritis present and indices of activity and chronicity would allow continuous monitoring of these parameters without the morbidity, cost, and inconvenience of renal biopsy. A number of proteins or protein modifications have been proposed as potential markers in lupus nephritis [912]. None have been shown to be useful to differentiate disease class, chronicity or activity. We have used two-dimensional gel electrophoresis and artificial neural networks to identify urinary protein spots that correlate with these clinical measurements.


Urine was collected from 20 consecutive patients prior to undergoing renal biopsy for evaluation of lupus nephritis at the Medical University of South Carolina according to a protocol approved by the Institutional Review Board. Generally, patients were newly diagnosed with proteinuria and were not on more than 10 mg of prednisone. Ten of the patients were taking disease-modifying agents at the time the urine sample was collected (hydroxychloroquine-7, azathioprine-1, mycophenolate mofetil-2). An experienced renal pathologist determined ISN/RPS classification of the disease and indices of activity and chronicity prior to analysis of the urine proteins. Cellular elements and debris were removed by centrifugation. Samples were frozen at −80°C until they were prepared for two-dimensional gel electrophoresis. After thawing, an equal volume of acetone was added to 10 mL of urine and incubated for 10 minutes. The 9000 × g precipitate was resuspended in 100 μL of a buffer containing 7 mol/L urea, 2 mol/L thiourea, 2% CHAPS and 1% ASB-14. One gel was run for each patient. A complete description of the urine sample preparation and two-dimensional gel electrophoresis protocol can be found in the MI2DG protocol database at (protocol name Precipitated_Urine_1). This publicly accessible data base is designed to describe all of the important variables in two-dimensional gel electrophoresis in sufficient detail to limit variability between laboratories.

Urine protein concentration was adjusted to 200 μg in 185 μL with a buffer containing 7 mol/L urea, 2 mol/L thiourea, 2% CHAPS, 1% ASB-14, 0.2% 3 to 10 ampholytes, and 50 mmol/L dithiothreitol (DTT). Two-dimensional electrophoresis was performed as previously described over a pH range of 4 to 7 and on 8% to 16% gradient sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels [13]. Gels were stained overnight with Sypro Ruby (Molecular Probes, Eugene, OR, USA) and imaged on an FX Pro Plus fluorescent imager (Bio-Rad, Hercules, CA, USA). Gel images were aligned using PDQuest software. Protein matches and intensities were converted to a Web-based open-source public infrastructure called annotated gel markup language (AGML) as previously described [14, 15]. The AGML formatted structure was processed by code written in MATLAB (MathWorks Inc), and spot intensities were ranked by intensity and expressed as quantiles [16]. An artificial neural network models was trained using the ranked protein intensities for 213 spots and three demographic parameters (gender, race, and age) by MATLAB code written along the guidelines previously proposed [17]. The artificial neural network fits a universal function to the input data using an iterative approach where each change in the function is evaluated to determine its effect on the accuracy of the outputs. The predictive value of each spot was evaluated by sensitivity analysis. Each seven artificial neural network classifiers were trained independently. Receiver operator characteristics (ROC) curves were determined to assess the accuracy of the results for ISN/RPS class and indices of activity and chronicity. The data from all 20 patients was then used to predict the classification of the training set. Sensitivity and specificity values were calculated for each diagnosis.


Urine was collected from twenty consecutive patients with SLE at the time of renal biopsy. There were five Caucasians, 15 African-Americans, two males and 18 females. Mean age was 27 ± 8 years. ISN/RPS class of lupus nephritis and indices of activity and chronicity were determined by a renal pathologist. Nine patients had histologic characteristics of more than one class of lupus nephritis and were assigned to more than one ISN/RPS class.

A representative example of a two-dimensional gel of urine proteins from a patient with class V lupus nephritis is shown in Figure 1. We aligned 213 protein spots across the gels. Analysis by artificial neural networks was done on protein abundance normalized by quantile. Output values were the binary value for each ISN/RPS class II, III, IV, and V and an ordinal value for activity and chronicity. ROC curves were derived for ISN/RPS classes. The area under the curve (AUC) for the ROC curve was greater than 0.85 for all classes. The r value for activity index was a relatively poor 0.77, but the r value for the chronicity index was much better at 0.87 (Table 1). ROC AUC and correlation coefficient are not reported for class II since there were only two positive diagnoses.

Fig. 1
Two-dimensional gel separation of proteins from a patient with class V lupus nephritis
Table 1
Statistics of sample classification by artificial neural networks

The original data set was analyzed by the trained artificial neural networks to determine its ability to predict the disease. For each gel, a prediction of presence or absence of each class of lupus nephritis was given. Sensitivity was 86% or greater for all classes. The sensitivity was lowest in class III disease, in which six of the seven cases were correctly identified. The specificity was 92% or greater for all classes except V, where only four of the eight patients who were negative for class V were correctly identified. Interestingly, all the false positive identifications of class V had class III or IV lupus present and were correctly identified as such. Clinically, a patient with both proliferative (class III or IV) and membranous (class V) would be treated for the more aggressive proliferative lesion, so the false positive identification of class V would not have affected the treatment of the patient.

In addition to being useful for identifying the ISN/RPS class of lupus nephritis, urine markers could be used to predict the duration and amount of renal injury from the disease. We have trained the artificial neural networks to correlate with the histologic score for chronicity and activity. A high degree of correlation was obtained for chronicity (r = 0.87) (Fig. 2), and a lesser degree of correlation was obtained for the activity index (r = 0.77). It is worth recalling that these values are cross-validated and that the artifical neural network classifiers selected correspond to the median performer of a set of artificial neural network models that rely on different resampled subsets of the available data [1]. Therefore, the results obtained are representative to the same extent that the data set itself is sufficiently representative.

Fig. 2
Predicted vs. observed values from a trained artificial neural network for chronicity index

Classification of patients was based on patterns of protein abundance. In order to derive a clinically useful test to predict class, activity, and chronicity of lupus nephritis, the identity of the proteins that provide the most sensitivity in the trained network needs to be determined. Both the amount of sensitivity for an individual protein in a given gel and the overall amount of sensitivity for the analysis was determined. Analysis of sensitivity for each of the outcomes was performed. Interestingly, most of the sensitivity was derived by a limited set of spots. Table 2 lists in order the ten spot numbers or demographic factors that provided the most sensitivity. Spot numbers 5, 77, and 44 were near the top for amount of sensitivity provided for most of the analyses. None of the spots alone could differentiate between classes. Race, gender, or age were important in the analysis for several diagnoses. Total amount of sensitivity provided by the top ten variables for each diagnosis is shown at the bottom of the table. Using matrix-assisted laser desorption-ionization tandem mass spectrometry (MALDI-TOF-TOF) and informatic tools we have identified the following proteins that provide the highest sensitivity: spot 5, α-1 acid glycoprotein; spot 44, α1 microglobulin; spot 52, zinc α-2 glycoprotein; spot 53, zinc α-2 glycoprotein; spot 75, IgG κ light chain; and spot 77, α1 microglobulin.

Table 2
Ranked sensitivities for diagnosis of lupus class, activity, and chronicity


Global protein quantitation strategies provide an opportunity to identify biomarkers that can identify diseases in a manner not possible with the candidate approach based on more limited numbers of proteins [18]. Quantification of urine proteins by two-dimensional gel electrophoresis was first done in the 1970s [19]. In spite of this, relatively little progress has been made to identify urinary biomarkers. Proteomic techniques have been used to look for urinary protein markers for stroke [20], bladder carcinoma [21], radiocontrast administration [22], and skeletal muscle toxicity [23]. Recently, surface-enhanced laser desorption and ionization (SELDI) has been used to identify urinary markers of acute renal allograft rejection [2426]. These studies have provided interesting information about urine protein profiles but have not yet described a set of proteins that can be easily adapted to create a clinically useful test.

The studies described here have demonstrated an approach to urinary biomarker identification that successfully identified protein spots that can serve as surrogates for a renal biopsy to identify class and chronicity index of disease in this cohort. The set of patients used in the analysis included patients who had more than one coexisting class of lupus nephritis. This set mimics patients that would be seen in the clinic for whom the diagnostic assay would be most useful. Evaluation of chronicity by urine test would also be very useful although as demonstrated in Figure 2, the correlation between predicted and observed is better in this set for nonzero values of chronicity. We have identified six of the protein spots as four unique proteins. These proteins are glycosylated serum proteins. Zinc α-2 glycoprotein and α1 microglobulin are both present in at least two charge forms among the proteins we have identified. These data suggest that differences in alterations in glomerular permeability to charged proteins may be partially responsible for the ability of the urine proteins to serve as biomarkers. The identity of the other proteins may reflect altered glomerular permeability to specific proteins, altered renal metabolism, or proteins that originate within the glomerulus, tubule, or interstitium or from inflammatory cells. The machine learning model was not able to predict activity index and has a false positive and false negative rate for the other diagnoses that would be unacceptable for a clinical assay. A larger numbers of patients will need to be studied in order to create models that can more reliably predict classes, chronicity, and activity. The current model will require external validation in a new set of patients. This validation should be done after antibodies are raised to the proteins that provide the most sensitivity and are validated. A clinically useful multiplexed enzyme-linked immunosorbent assay (ELISA)-type assay could be developed based on antibodies generated against these proteins and would be more sensitive, reproducible, simple and rapid than two-dimensional gel electrophoresis.


These studies were supported by the MUSC GCRC (RR01070) and Department of Veterans Affairs and grants from Dialysis Clinics, Inc., and the NIH (R21AR051719). Data analysis was supported with Federal funds as part of the NHLBI Proteomics Initiative from the National Heart, Lung, and Blood Institute, National Institutes of Health, under Contract No. N01-HV-28181.


1. U.S. Renal Data System. USRDS 2004 Annual Data Report: Atlas of End-Stage Renal Disease in the United States. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 2004.
2. Weening JJ, D’Agati VD, Schwartz MM, et al. The classification of glomerulonephritis in systemic lupus erythematosus revisited. Kidney Int. 2004;65:521–530. [PubMed]
3. Bajaj S, Albert L, Gladman DD, et al. Serial renal biopsy in systemic lupus erythematosus. J Rheumatol. 2000;27:2822–2826. [PubMed]
4. Rush PJ, Baumal R, Shore A, et al. Correlation of renal histology with outcome in children with lupus nephritis. Kidney Int. 1986;29:1066–1071. [PubMed]
5. Korbet SM, Lewis EJ, Schwartz MM, et al. Factors predictive of outcome in severe lupus nephritis. Lupus Nephritis Collaborative Study Group. Am J Kidney Dis. 2000;35:904–914. [PubMed]
6. Moroni G, Pasquali S, Quaglini S, et al. Clinical and prognostic value of serial renal biopsies in lupus nephritis. Am J Kidney Dis. 1999;34:530–539. [PubMed]
7. Gadgil NM, Ranadive NU, Shete M, Bhatia S. Clinical, morphological, immunological correlation of kidney biopsies and prognostication. Indian J Pathol Microbiol. 1999;42:3–10. [PubMed]
8. Najafi CC, Korbet SM, Lewis EJ, et al. Significance of histologic patterns of glomerular injury upon long-term prognosis in severe lupus glomerulonephritis. Kidney Int. 2001;59:2156–2163. [PubMed]
9. Molad Y, Miroshnik E, Sulkes J, et al. Urinary soluble VCAM-1 in systemic lupus erythematosus: A clinical marker for monitoring disease activity and damage. Clin Exp Rheumatol. 2002;20:403–406. [PubMed]
10. De Muro P, Faedda R, Formato M, et al. Urinary glycosamino-glycans in patients with systemic lupus erythematosus. Clin Exp Rheumatol. 2001;19:125–130. [PubMed]
11. Bobkova I, Lysenko L, Polyantseva L, Tareyeva I. Urinary von Willebrand factor as a marker of lupus nephritis progression. Nephron. 2001;87:369–370. [PubMed]
12. Kipen Y, Will R, Strauss BJ, Morand EF. Urinary excretion of the pyridinium cross-links of collagen in systemic lupus erythematosus. Clin Rheumatol. 1998;17:271–276. [PubMed]
13. Lefler DM, Pafford RG, Black NA, et al. Identification of proteins in slow continuous ultrafiltrate by reversed-phase chromatography and proteomics. J Proteome Res. 2004;3:1254–1260. [PubMed]
14. Stanislaus R, Jiang LH, Swartz M, et al. An XML standard for the dissemination of annotated 2D gel electrophoresis data complemented with mass spectrometry results. BMC Bioinformatics. 2004;5:9. [PMC free article] [PubMed]
15. Stanislaus R, Chen C, Franklin J, et al. AGML central: web based gel proteomic infrastructure. Bioinformatics. 2005;21:1754–1757. [PubMed]
16. Almeida JS, Stanislaus R, Krug EL, Arthur JM. Normalization and analysis of residual variation in 2D gel electrophoresis for quantitative differential proteomics. Proteomics. 2005;5:1242–1249. [PubMed]
17. Almeida JS. Predictive non-linear modeling of complex data by artificial neural networks. Curr Opin Biotechnol. 2002;13:72–76. [PubMed]
18. Anderson NL, Anderson NG. The human plasma proteome: History, character, and diagnostic prospects. Mol Cell Proteomics. 2002;1:845–867. [PubMed]
19. Anderson NG, Anderson NL, Tollaksen SL. Proteins of human urine: I. Concentration and analysis by two-dimensional electrophoresis. Clin Chem. 1979;25:1199–1210. [PubMed]
20. Sironi L, Tremoli E, Miller I, et al. Acute-phase proteins before cerebral ischemia in stroke-prone rats: Identification by proteomics. Stroke. 2001;32:753–760. [PubMed]
21. Celis JE, Wolf H, Ostergaard M. Bladder squamous cell carcinoma biomarkers derived from proteomics. Electrophoresis. 2000;21:2115–2121. [PubMed]
22. Hampel DJ, Sansome C, Sha M, et al. Toward proteomics in uroscopy: Urinary protein profiles after radiocontrast medium administration. J Am Soc Nephrol. 2001;12:1026–1035. [PubMed]
23. Dare TO, Davies HA, Turton JA, et al. Application of surface-enhanced laser desorption/ionization technology to the detection and identification of urinary parvalbumin-alpha: A biomarker of compound-induced skeletal muscle toxicity in the rat. Electrophoresis. 2002;23:3241–3251. [PubMed]
24. Clarke W, Silverman BC, Zhang Z, et al. Characterization of renal allograft rejection by urinary proteomic analysis. Ann Surg. 2003;237:660–664. [PubMed]
25. O’Riordan E, Orlova TN, Mei J, et al. Bioinformatic analysis of the urine proteome of acute allograft rejection. J Am Soc Nephrol. 2004;15:3240–3248. [PubMed]
26. Schaub S, Rush D, Wilkins J, et al. Proteomic-based detection of urine proteins associated with acute renal allograft rejection. J Am Soc Nephrol. 2004;15:219–227. [PubMed]