Lupus (2018) 0, 1–16

journals.sagepub.com/home/lup

PAPER

Urinary vitamin D-binding protein, a novel biomarker for lupus

nephritis, predicts the development of proteinuric flare
DJ Go1,2,* , JY Lee2,*, MJ Kang2, EY Lee3, EB Lee3, EC Yi2 and YW Song2,3
1
Division of Rheumatology, Department of Internal Medicine, Hallym University Kangnam Sacred Heart Hospital, Seoul, Korea; 2Department of
Molecular Medicine and Biopharmaceutical Sciences, Graduate School of Convergence Science and Technology, and College of Medicine, Medical
Research Institute, Seoul National University, Seoul, Korea; and 3Division of Rheumatology, Department of Internal Medicine, Seoul National
University Hospital, Seoul, Korea

Lupus nephritis (LN) is a major complication of systemic lupus erythematosus (SLE).

Conventional biomarkers for assessing renal disease activity are imperfect in predicting clinical
outcomes associated with LN. The aim of this study is to identify urinary protein biomarkers
that reliably reflect the disease activity or predict clinical outcomes. A quantitative proteomic
analysis was performed to identify protein biomarker candidates that can differentiate
between SLE patients with and without LN. Selected biomarker candidates were further
verified by enzyme-linked immunosorbent assay using urine samples from a larger cohort
of SLE patients (n ¼ 121) to investigate their predictive values for LN activity measure.
Furthermore, the association between urinary levels of a selected panel of potential
biomarkers and prognosis of LN was assessed with a four-year follow-up study of renal
outcomes. Urinary vitamin D-binding protein (VDBP), transthyretin (TTR), retinol binding
protein 4 (RBP4), and prostaglandin D synthase (PTGDS) were significantly elevated in SLE
patients with LN, especially in patients with active LN (n ¼ 21). Among them, VDBP well
correlated with severity of proteinuria (rho ¼ 0.661, p < 0.001) and renal SLE Disease Activity
Index (renal SLEDAI) (rho ¼ 0.520, p < 0.001). In the four-year follow-up, VDBP was a sig-
nificant risk factor (hazard ratio 9.627, 95% confidence interval 1.698 to 54.571, p ¼ 0.011) for
the development of proteinuric flare in SLE patients without proteinuria (n ¼ 100) after adjust-
ments for multiple confounders. Urinary VDBP correlated with proteinuria and renal
SLEDAI, and predicted the development of proteinuria. Lupus (2018) 0, 1–16.

Key words: Systemic lupus erythematosus; lupus nephritis; vitamin D-binding protein; urinary
biomarker

Introduction
morbidity and mortality of SLE patients,2,3 early
diagnosis and subsequent immunosuppressive
Systemic lupus erythematosus (SLE) is a chronic therapy are crucial to prevent progression to
autoimmune disease that can cause multi-organ renal insufficiency.4,5 Renal biopsy is often
failure. Approximately 50% of patients with SLE required to confirm the diagnosis of LN and to
develop clinical renal disease.1 Given that lupus determine the treatment strategy according to
nephritis (LN) contributes to the increased histologic features. However, it is rather invasive,
occasionally leading to procedure-related compli-
*D.J.G. and J.Y.L. contributed equally to this study. cations like bleeding, pain, and arteriovenous fis-
Correspondences to: Yeong Wook Song (about clinical interpretation), tula.6,7 Although anti-double-stranded DNA
Division of Rheumatology, Department of Internal Medicine, Seoul (anti-dsDNA) antibodies and complement com-
National University Hospital, 101, Daehak-ro, Jongno-gu, Seoul, ponent 3 and 4 (C3 and C4) levels, as serologic
03080, Korea. biomarkers for systemic disease, are associated
Email: ysong@snu.ac.kr
Eugene C. Yi (about proteomic analysis), Department of Molecular
with the presence of LN, they have insufficient
Medicine and Biopharmaceutical Sciences, Seoul National University, sensitivity and specificity to reflect disease activity
103, Daehak-ro, Jongno-gu, Seoul, 03080, Korea. of LN.8 Therefore, there is a need for noninvasive
Email: euyi@snu.ac.kr biomarkers that can provide accurate information
Received 21 November 2017; accepted 1 May 2018 about disease activity of LN.
! The Author(s), 2018. Reprints and permissions: http://www.sagepub.co.uk/journalsPermissions.nav 10.1177/0961203318778774
 Urinary VDBP in lupus nephritis
DJ Go et al.
2
Urine is an ideal biological sample for detection
of LN biomarkers since it is easy to harvest, and
urine proteins may be more specific for renal
inflammation than serum proteins.9 Several urinary
protein biomarkers have been reported as potential
predictors of renal disease activity of SLE. These
include neutrophil gelatinase-associated lipocalin
(NGAL), monocyte chemoattractant protein-1
(MCP-1), tumor necrosis factor-like weak inducer
of apoptosis (TWEAK), and prostaglandin D syn-
thase (PTGDS).10–15 However, the prognostic
values of these biomarkers have not yet been fully
evaluated for predicting LN flares.
Recent advancements in mass spectrometry
(MS)-based proteomics have facilitated the highly
sensitive measurement of urinary biomarker candi-
dates for LN.16 Here, we used a global proteomic
profiling experiment to identify potential bio-
markers for predicting LN using pooled urine
samples of SLE patients with or without LN.
Moreover, a panel of candidate urine proteins
was further quantified in SLE patients. We then
investigated the association between the selected
biomarker candidates and prognosis of LN in
regard to response or flare of proteinuria.
Patients and methods
SLE patients with and without LN
This study was conducted as a cross-sectional study
in measuring urine biomarker concentration and
evaluating the association with LN status in SLE
patients. Urine was collected once from two sets of
independent cohorts (screening and validation) in
the Rheumatology Clinic of Seoul National
University Hospital (SNUH) within two months
(February and March 2012). One cohort comprised
eight SLE patients who fulfilled the 1997 American
College of Rheumatology (ACR) criteria17 with
(n ¼ 4) and without (n ¼ 4) LN for proteomics
screening of biomarker candidates. The other
cohort comprised 121 SLE patients for validation
of biomarker candidates by enzyme-linked
immunosorbent assay (ELISA). LN was classified
according to the 1997 ACR criteria for renal
involvement,17 and biopsy-proven LN was further
classified according to the World Health
Organization classification.18 SLE patients were
divided into three groups. The first group was
patients with SLE without LN (urine protein-
creatinine ratio (UPCR) < 0.5 mg/mg without pre-
vious history of active LN). The second group was
patients with inactive LN (UPCR < 0.5 mg/mg with
Lupus
previous history of active LN). In these patients,
urine samples were collected after 50.2  44.5
(mean  standard deviation) months from recovery
of active LN. The third group was patients with
active LN (renal SLE Disease Activity Index
(SLEDAI)  419 with random UPCR  0.5 mg/mg,
equivalent to 24-hour urine protein  500 mg). The
demographic and clinical characteristics of each
group including the laboratory findings within
one month of urine sampling date (baseline) were
obtained.
Additionally, patients’ medical records were
reviewed to examine the further occurrences of
remission or flare of proteinuria after initial
urine sampling for approximately four years.
Remission of proteinuria was defined as return
to random UPCR < 0.5 mg/mg in SLE patients
with active LN (UPCR  0.5 mg/mg).20 The defin-
ition of flare was the development of significant
proteinuria, defined by UPCR  1.0 mg/mg or 24-
hour urine protein  1.0 g, in SLE patients with-
out definite proteinuria (UPCR < 0.5 mg/mg).21,22
This study was approved by the Institutional
Review Board of SNUH (IRB No. 1208-086-
422) and informed consent was obtained from
the study participants.
Protein extraction from urine samples
Pooled urine samples (5 ml aliquots) from patients
were analyzed following the general urinary prote-
omic method.23 Urinary proteins were extracted
using methanol precipitation and the protein pellets
were resuspended with 5 mM ethylenedinitrilo tet-
raacetic acid (EDTA) disodium salt in 50 mM Tris/
HCl (pH 7.5) including protease inhibitor. The pro-
tein amount was measured using the Micro BCA
Protein Assay Kit (Thermo Scientific, Rockford,
IL, USA). Then, albumin was depleted from the
extracted proteins (1 mg each) using an Affinity
Removal Spin Cartridge (Agilent Technologies,
Santa Clara, CA, USA). Following the manufac-
turer’s protocol, urinary protein samples were
diluted three-fold with depletion Buffer A (Agilent
Technologies) and filtered through a 0.22 mm spin
filter (Agilent Technologies). The diluted urinary
protein samples were loaded onto the Spin
Cartridge and centrifuged at 100 g for 1.5 minutes.
The flow-through fraction was collected and con-
centrated using an Amicon 3 kDa molecular weight
cutoff filter (Millipore, Bedford, MA, USA).
Depleted urinary protein concentration was mea-
sured again using the Micro BCA Protein Assay
Kit (Thermo Scientific).

Sodium dodecyl sulfate-polyacrylamide gel
electrophoresis fractionation and in-gel digestion
Equal amounts of albumin-depleted proteins
(100 mg) were separated on a 4%12% Bis-Tris
NuPAGE gel (Novex; Thermo Scientific),
and stained with Coomassie Brilliant Blue
(Sigma-Aldrich, St Louis, MO, USA). The entire
gel lane was cut into 12 pieces and subjected
to in-gel tryptic digestion following the general
protocol.24 Briefly, excised protein bands were
destained, reduced with 20 mM dithiothreitol, and
alkylated with 55 mM iodoacetamide. After dehy-
dration with acetonitrile (ACN), the proteins were
digested with 12.5 ng/ml modified trypsin (Promega,
Madison, WI, USA) in 50 mM ammonium bicar-
bonate overnight at 37 C. Peptides were extracted
from the gel slices with 50% (v/v) ACN in 0.1%
(v/v) formic acid (FA). The eluates were dried using
a Centrivap concentrator (Labconco, Kansas City,
MO, USA) and stored at 20 C until use.
MS analysis and database search
Extracted peptide samples were suspended in solvent
A (0.1% FA in water, Fisher Scientific, Pittsburgh,
PA, USA), loaded onto an EASY-Spray C18
column (75 mm50 cm, 2 mm) and separated with a
2%–35% gradient of solvent B (0.1% FA in ACN)
for 65 minutes at a flow rate of 300 nl/minute. MS
spectra were recorded on a Q-Exactive hybrid quad-
rupole-Orbitrap MS (Thermo Fisher Scientific,
San Jose, CA, USA) interfaced with a nano-ultra-
high-performance liquid chromatography system
(Easy-nLC1000; Thermo Scientific). Collected MS/
MS raw files were converted to mzXML files using
the Trans-Proteomic Pipeline (version 4.4) and ana-
lyzed using the Sequest (version 27) algorithm in the
SORCERER (Sage-N Research, Milpitas, CA,
USA) platform. The search was performed using
the International Protein Index human database
(version 3.83, 186,578 entries). Full tryptic specificity
and up to two missed cleavage sites were allowed.
Mass tolerances for precursor ions and fragment
ions were set to 10 ppm and 1 Da, respectively.
Fixed modification for carbamidomethyl-cysteine
and variable modifications for methionine oxida-
tion were used. All proteins with a Protein
Prophet probability of  99% with minimum two
peptides and a PeptideProphet probability
of  95% (false discovery rate (FDR)  0.15%)
were identified.25 Relative protein quantitation
was accomplished using spectral counting. Liquid
chromatography tandem MS (LC-MS/MS) data
were normalized to compare the abundances of
proteins between samples using Scaffold version
Urinary VDBP in lupus nephritis
DJ Go et al.
3
4.5.3 software (Proteome Software, Portland, OR,
USA). The normalized spectral counts from dupli-
cate runs were compared using the R program with
power law global error model (PLGEM) software
(http://www.bioconductor.org) to identify differen-
tially expressed proteins (DEPs).26
ELISA
Potential biomarker candidates selected from the
global proteomic profiling experiment were again
quantified by ELISA. The urinary concentrations
of vitamin D-binding protein (VDBP) and trans-
thyretin (TTR) were measured by sandwich
ELISA methods, using mouse monoclonal anti-
VDBP antibody (Abcam, Cambridge, MA, USA)
and mouse monoclonal anti-TTR antibody
(Abcam), respectively. An indirect ELISA was per-
formed to measure the urinary prostaglandin D
synthase (PTGDS) concentration using antibody
solution of pre-mixed rat monoclonal anti-
PTGDS antibody (Cayman Chemical, Ann
Arbor, MI, USA). Urinary retinol-binding protein
4 (RBP4) concentrations were measured with
human RBP4 ELISA kits (R&D Systems,
Minneapolis, MN, USA; intra-assay coefficient of
variation (CV) 5.7%–8.1% and interassay CV
5.8%–8.6%) in accordance with the manufacturer’s
instruction. Concentrations of urinary biomarkers
were derived from an average of the duplicated
assay for one sample and standardized to creatinine
(Cr) concentrations in the same urine. Assay preci-
sion in terms of intra- and interassay CVs for
means of duplicate determinations were less than
10% and 15% respectively for the measurements.
A researcher who performed ELISA was blinded to
the clinical data.
Statistical analyses
Mann-Whitney U test (median (interquartile
range)) or Student’s t test (mean  standard devi-
ation) were used to compare continuous variables,
and chi-square or Fisher exact test were used when
comparing categorical variables, as appropriate.
The relationships between urinary biomarkers and
laboratory findings in LN were assessed with
Spearman’s rank correlation test. Receiver operat-
ing characteristic (ROC) curves were used to calcu-
late the area under the curve (AUC) for urinary
biomarkers and to determine the cutoff values for
prognostic prediction. Logistic regression and Cox
proportional hazard regression analyses were used
to predict the occurrence of remission and flare
according to multiple covariates, respectively.
Kaplan-Meier curves and log-rank tests were
Lupus
 Urinary VDBP in lupus nephritis
DJ Go et al.
4
implemented to estimate the proteinuria-free sur-
vival rate with respect to low vs. high level of urin-
ary biomarkers. A p value < 0.05 was considered
statistically significant. SPSS software version 22.0
(SPSS Inc, Chicago, IL, USA) was utilized for data
analysis.
Results
Urinary proteome profiling for identification of
biomarker candidates
To identify potential urinary protein biomarkers
for differentiating between SLE patients with and
without LN, quantitative urinary proteome profil-
ing of pooled urine samples was performed in
patients with and without LN. SLE patients with
LN (n ¼ 4) had significantly shorter duration of
SLE and lower C3 level than those without LN
(n ¼ 4) in this screening analysis. Three patients
with LN received high-dose corticosteroids (CS)
(prednisolone  0.5 mg/kg) and one of them
received cyclophosphamide (CYC) pulse therapy
at the time of urine sampling (Supplementary file
1: Table S1). After duplicated LC-MS/MS analysis,
we merged the protein database search results and
identified 1373 proteins (protein probabil-
ity  99.0%, minimum two peptides, peptide prob-
ability  95.0%, peptide FDR  0.15%). Label-free
quantitative proteome analysis based on spectral
counting of identified proteins using PLGEM soft-
ware revealed 153 DEPs (p  0.001). Fifty-six pro-
teins were upregulated and 97 proteins were
downregulated in patients with LN compared to
those without LN (Table 1).
Considering the extensive level (p < 0.0002) of
divergence between SLE with LN and without
LN, VDBP (signal-to-noise (STN) ¼ 34.5,
p < 0.0002), TTR (STN ¼ 38.6, p < 0.0002), and
RBP4 (STN ¼ 104.6, p < 0.0002) were selected for
further evaluation using ELISA assay. The urinary
level of PTGDS (STN ¼ 9.6, p ¼ 0.001) was
downregulated in the SLE patients with LN com-
pared to those without LN in the quantitative
proteome analysis. This contradicted the known
expression in the urine of LN patients.27,28
Nonetheless, PTGDS was included as a reference
in the ELISA-based verification experiment.
Baseline characteristics of SLE patients in a
validation study
The demographic and clinical characteristics of
patients (n ¼ 121) included in ELISA study are
Lupus
summarized in Table 2. The mean age of the
SLE patients was 41.9  12.6 years and 14
(11.6%) were male. LN had been diagnosed
according to the 1997 ACR criteria in 62
(51.2%) patients, including 46 patients with
biopsy-proven LN. Patients with LN had a
higher proportion of being male, lower estimated
glomerular filtration rate (eGFR) assessed using
the Modification of Diet in Renal Disease for-
mula,29 and higher SLEDAI than did those with-
out LN. Comparing with inactive LN, active LN
(n ¼ 21) was associated with younger age, short
disease duration of SLE or LN, higher anti-
dsDNA titer, and lower C3 level. Patients with
active LN received higher daily doses of CS and
more frequent angiotensin-converting enzyme
(ACE) inhibitor or angiotensin receptor blocker
(ARB) medications.
Levels of urinary biomarkers according to LN status
ELISA analysis revealed higher urine concentra-
tions for each of the biomarker candidates in the
patients with active LN (n ¼ 21) compared with
those with inactive LN (n ¼ 41): VDBP median
211.36 (interquartile range 15.19–609.71) vs. 2.68
(0.61–13.76) ng/mgCr, TTR 69.48 (3.39–308.57)
vs. 0.54 (0.08–11.42) ng/mgCr, RBP4 1086.80
(164.46–5859.26) vs. 90.72 (62.70–339.47) ng/
mgCr, and PTGDS 2306.84 (961.45–19336.52) vs.
701.22 (340.02–1250.76) ng/mgCr (Figure 1).
Interestingly, urine levels of biomarker candidates
in patients with inactive LN (n ¼ 41), except VDBP,
were also significantly higher compared with those
in SLE patients without LN (n ¼ 59): TTR 0.54
(0.08–11.42) vs. 0.13 (0.01–4.53) ng/mgCr, RBP4
90.72 (62.70–339.47) vs. 55.27 (12.68–129.92) ng/
mgCr, and PTGDS 701.22 (340.02–1250.76) vs.
335.60 (132.32–709.07) ng/mgCr (Figure 1). When
we analyzed patients with biopsy-proven LN
(n ¼ 46) only, the comparison results of urinary
biomarker levels according to LN status
were similar to the aforementioned findings
(Supplementary file 2: Figure S1). However,
there was no difference in urine concentrations of
all biomarker candidates according to drug use
(mean daily CS doses, presence of vitamin D sup-
plement, bisphosphonates, and ACE inhibitor or
ARB during one week prior to sample collection)
in each patient group including active LN.
In patients with inactive LN, urine levels of all
biomarker candidates were not correlated with
disease duration or periods from the recovery of
active LN.
 Urinary VDBP in lupus nephritis
DJ Go et al.
5

Table 1 Differentially expressed 153 urinary proteins in SLE patients with LN by proteomics analysis (p value  0.001)
Spectral count

SLE without LN SLE with LN PLGEM

Accession First Second First Second

No number Gene symbols Protein names MW run run run run STN p value

1 IPI00553177 SERPINA1 Isoform 1 of Alpha-1-antitrypsin 47 kDa 1093 1101 3294 3246 156.6 0.0000
2 IPI00022463 TF Serotransferrin 77 kDa 1789 1754 3689 3648 152.3 0.0000
3 IPI00022420 RBP4 Retinol-binding protein 4 23 kDa 227 217 1055 1079 104.6 0.0000
4 IPI00478003 A2M Alpha-2-macroglobulin 163 kDa 295 308 932 929 75.8 0.0000
5 IPI00017601 CP Ceruloplasmin 122 kDa 668 655 1385 1363 61.1 0.0000
6 IPI00022429 ORM1 Alpha-1-acid glycoprotein 1 24 kDa 321 314 821 868 59.0 0.0000
7 IPI00784817  Anti-RhD monoclonal T125 gamma1 52 kDa 0 0 8 2 57.3 0.0000
heavy chain
8 IPI00022229 APOB Apolipoprotein B-100 516 kDa 30 38 292 283 55.0 0.0000
9 IPI00783987 C3 Complement C3 (Fragment) 187 kDa 316 314 882 822 50.0 0.0000
10 IPI00004656 B2M Beta-2-microglobulin 14 kDa 85 85 267 300 49.1 0.0000
11 IPI00166729 AZGP1 Zinc-alpha-2-glycoprotein 34 kDa 491 483 833 852 42.3 0.0000
12 IPI00021841 APOA1 Apolipoprotein A-I 31 kDa 121 117 428 439 41.3 0.0000
13 IPI00020091 ORM2 Alpha-1-acid glycoprotein 2 24 kDa 138 149 338 330 41.3 0.0000
14 IPI00032179 SERPINC1 Antithrombin-III 53 kDa 186 176 477 496 40.2 0.0000
15 IPI00887169 IGLV1-44 Putative uncharacterized protein 25 kDa 0 0 11 11 39.8 0.0000
16 IPI00855916  Transthyretin 20 kDa 97 93 347 361 38.2 0.0000
17 IPI00555812 GC Vitamin D-binding protein 53 kDa 150 150 415 430 34.5 0.0000
18 IPI00022895 A1BG Isoform 1 of Alpha-1B-glycoprotein 54 kDa 364 362 670 674 30.5 0.0000
19 IPI00550991 SERPINA3 cDNA FLJ35730 fis, clone 51 kDa 350 327 646 646 30.0 0.0000
TESTI2003131, highly similar
to ALPHA-1-ANTICHYMOTRYPSIN
20 IPI00399007 IGHG2 Putative uncharacterized protein 46 kDa 1325 1285 1568 1580 29.1 0.0000
DKFZp686I04196 (Fragment)
21 IPI00022488 HPX Hemopexin 52 kDa 268 253 559 544 28.8 0.0000
22 IPI00025426 PZP Isoform 1 of Pregnancy zone protein 164 kDa 0 0 37 37 28.7 0.0000
23 IPI00784842  Putative uncharacterized protein 52 kDa 68 71 89 92 27.7 0.0000
DKFZp686G11190
24 IPI00305461 ITIH2 Inter-alpha (Globulin) inhibitor H2, 107 kDa 35 45 233 227 27.5 0.0000
isoform CRA_a
25 IPI00785084  IGH@ protein 51 kDa 535 523 712 706 27.1 0.0000
26 IPI00215746 FABP4 Fatty acid-binding protein, adipocyte 15 kDa 5 5 113 116 27.0 0.0000
27 IPI00744476  IGL@ protein 25 kDa 2 0 22 11 26.9 0.0000
28 IPI00418153 IGHM Putative uncharacterized 57 kDa 24 24 41 40 24.5 0.0002
protein DKFZp686I15212
29 IPI00019943 AFM Afamin 69 kDa 119 121 270 277 24.2 0.0002
30 IPI00645363  Putative uncharacterized protein 52 kDa 82 73 116 109 21.3 0.0002
DKFZp686P15220
31 IPI00892604  Complement component C4B 193 kDa 214 224 337 358 20.5 0.0002
(Childo blood group) 2
32 IPI00449920 IGHA1 cDNA FLJ90170 fis, 53 kDa 25 25 44 36 19.8 0.0003
clone MAMMA1000370,
highly similar to Igalpha-1 chain
C region
33 IPI00022426 AMBP Protein AMBP 39 kDa 1518 1525 1908 1898 19.1 0.0003
34 IPI01014563 FTL Ferritin light chain 20 kDa 9 8 75 72 18.9 0.0003
35 IPI00215983 CA1 Carbonic anhydrase 1 29 kDa 60 54 227 219 18.8 0.0003
36 IPI00009197 REG1B Lithostathine-1-beta 19 kDa 0 0 11 15 18.1 0.0003
37 IPI00019580 PLG Plasminogen 91 kDa 316 325 439 425 18.0 0.0003
38 IPI00019568 F2 Prothrombin (Fragment) 70 kDa 79 74 192 179 17.5 0.0003
39 IPI00304273 APOA4 Apolipoprotein A-IV 45 kDa 67 70 183 183 17.2 0.0003
40 IPI00217493 MB Myoglobin 17 kDa 8 10 71 68 15.8 0.0003
41 IPI00028413 ITIH3 Isoform 1 of Inter-alpha-trypsin 100 kDa 10 8 94 97 14.9 0.0004
inhibitor heavy chain H3
42 IPI00876888  cDNA FLJ78387 52 kDa 13 14 17 16 14.4 0.0005
43 IPI00654755 HBB Hemoglobin subunit beta 16 kDa 102 104 170 156 13.5 0.0007
(continued)

Lupus
 Urinary VDBP in lupus nephritis
DJ Go et al.
6
Table 1 Continued
Spectral count

SLE without LN SLE with LN PLGEM

Accession First Second First Second

No number Gene symbols Protein names MW run run run run STN p value

44 IPI00930442 IGHG4 Putative uncharacterized protein 52 kDa 14 20 55 50 12.9 0.0007

DKFZp686M24218
45 IPI00022417 LRG1 Leucine-rich alpha-2-glycoprotein 38 kDa 99 100 178 183 12.8 0.0007
46 IPI00829640 IGLV3-19 IGL@ protein 25 kDa 96 102 146 153 12.8 0.0007
47 IPI00006705 SCGB1A1 Uteroglobin 10 kDa 14 15 42 44 12.6 0.0007
48 IPI00171196 KRT13 Isoform 3 of Keratin, type I cytoskeletal 13 46 kDa 39 39 88 91 12.6 0.0007
49 IPI00154742  IGL@ protein 25 kDa 14 10 37 40 12.5 0.0007
50 IPI00827875  Lambda-chain 25 kDa 509 512 710 728 11.5 0.0008
51 IPI01013699 ITIH1 cDNA FLJ56821, highly similar to 76 kDa 22 15 78 71 11.1 0.0008
Inter-alpha-trypsin inhibitor
heavy chain
52 IPI00473011 HBD Hemoglobin subunit delta 16 kDa 4 5 10 7 10.7 0.0009
53 IPI00032220 AGT Angiotensinogen 53 kDa 72 78 148 152 10.5 0.0010
54 IPI00216691 PFN1 Profilin-1 15 kDa 15 15 47 57 10.5 0.0010
55 IPI00642632 IGLC7 Ig lambda-7 chain C region 11 kDa 3 0 10 6 10.0 0.0010
56 IPI00010290 FABP1 FABP1 protein (Fragment) 15 kDa 20 18 52 59 9.9 0.0010
57 IPI00009793 C1RL Complement C1r subcomponent-like protein 53 kDa 112 112 30 32 11.1 0.0010
58 IPI00021434 PI3 Elafin 12 kDa 26 25 0 0 11.1 0.0010
59 IPI00736755  Pepsin A preproprotein 42 kDa 102 106 22 24 11.3 0.0010
60 IPI00940136 CR1 Complement receptor type 1 224 kDa 43 54 2 2 11.5 0.0010
61 IPI00376403 SPINT1 Isoform 1 of Kunitz-type protease inhibitor 1 58 kDa 59 62 3 3 11.6 0.0010
62 IPI00844511 MAN1A1 Mannosyl-oligosaccharide 73 kDa 71 63 3 0 11.6 0.0010
1,2-alpha-mannosidase IA
63 IPI00219426 PVR Isoform Gamma of Poliovirus 39 kDa 58 64 3 2 11.7 0.0010
receptor
64 IPI00290085 CDH2 Cadherin-2 100 kDa 66 74 6 2 11.7 0.0010
65 IPI00043992 PVRL4 Poliovirus receptor-related protein 4 55 kDa 70 69 9 8 11.7 0.0010
66 IPI00218407 ALDOB Fructose-bisphosphate aldolase B 39 kDa 74 67 8 7 11.7 0.0010
67 IPI00216914 VMO1 Vitelline membrane outer layer 22 kDa 112 107 36 42 11.8 0.0010
protein 1 homolog
68 IPI00871227 HMCN1 Isoform 1 of Hemicentin-1 613 kDa 74 75 11 11 11.8 0.0010
69 IPI00176427 CADM4 Cell adhesion molecule 4 43 kDa 74 67 2 0 11.8 0.0010
70 IPI00414717 GLG1 Isoform 2 of Golgi apparatus protein 1 137 kDa 66 65 2 2 12.0 0.0010
71 IPI00306322 COL4A2 Collagen alpha-2(IV) chain 168 kDa 57 65 2 0 12.2 0.0010
72 IPI00289819 IGF2R Cation-independent 274 kDa 101 102 22 19 12.2 0.0010
mannose-6-phosphate receptor
73 IPI00006608 APP Isoform APP770 of Amyloid beta A4 87 kDa 56 61 4 2 12.3 0.0010
protein (Fragment)
74 IPI00332271 PTPRS Isoform PTPS-MEA of 216 kDa 80 85 11 9 12.3 0.0010
Receptor-type tyrosine-protein
phosphatase S
75 IPI00027235 ATRN Isoform 1 of Attractin 159 kDa 171 181 68 66 12.6 0.0009
76 IPI00008787 NAGLU Alpha-N-acetylglucosaminidase 82 kDa 104 104 11 13 12.6 0.0009
77 IPI00009028 CLEC3B Tetranectin 23 kDa 80 93 17 22 12.6 0.0009
78 IPI00003807 ACP2 Lysosomal acid phosphatase 48 kDa 68 71 0 0 12.7 0.0009
79 IPI00021085 GOLM1 Golgi membrane protein 1 46 kDa 66 68 2 2 12.8 0.0009
80 IPI00021085 PGLYRP1 Peptidoglycan recognition protein 1 22 kDa 121 112 31 41 12.9 0.0009
81 IPI00022620 SLURP1 Secreted Ly-6/uPAR-related 11 kDa 59 66 18 13 12.9 0.0009
protein 1
82 IPI00549330  immunoglobulin kappa locus-like 17 kDa 12 16 15 11 13.1 0.0009
83 IPI00247063 MME Neprilysin 86 kDa 60 75 0 0 13.2 0.0009
84 IPI00910975 CD300LG CMRF35-like molecule 9 isoform 30 kDa 94 88 11 11 13.2 0.0009
3 precursor
85 IPI00029751 CNTN1 Isoform 1 of Contactin-1 113 kDa 98 99 11 13 13.3 0.0009
86 IPI00027444 SERPINB1 Leukocyte elastase inhibitor 43 kDa 75 86 15 14 13.3 0.0009
87 IPI00022822 COL18A1 Isoform 2 of Collagen alpha-1(XVIII) chain 154 kDa 76 78 9 5 13.3 0.0009
88 IPI00441498 FOLR1 Folate receptor alpha 30 kDa 106 99 14 16 13.4 0.0009
89 IPI00298828 APOH Beta-2-glycoprotein 1 38 kDa 175 165 51 53 13.4 0.0009
(continued)

Lupus
 Urinary VDBP in lupus nephritis
DJ Go et al.
7
Table 1 Continued
Spectral count

SLE without LN SLE with LN PLGEM

Accession First Second First Second

No number Gene symbols Protein names MW run run run run STN p value

90 IPI00099670 CEL bile salt-activated lipase precursor 80 kDa 86 80 6 5 13.4 0.0009

91 IPI00006971 CD248 Isoform 1 of Endosialin 81 kDa 121 129 27 34 13.6 0.0009
92 IPI00021302 SUSD2 Sushi domain-containing protein 2 90 kDa 76 75 0 0 13.7 0.0009
93 IPI00827637  K light chain variable region (Fragment) 12 kDa 15 12 0 0 13.8 0.0009
94 IPI01013090 PLG cDNA FLJ58778, highly similar to 45 kDa 2 3 0 0 13.9 0.0008
Plasminogen
95 IPI00291488 WFDC2 Isoform 1 of WAP four-disulfide core 13 kDa 141 135 49 52 13.9 0.0008
domain protein 2
96 IPI00029168 LPA Apolipoprotein(a) 501 kDa 109 96 10 12 14.0 0.0008
97 IPI00003590 QSOX1 Isoform 1 of Sulfhydryl oxidase 1 83 kDa 79 72 0 3 14.3 0.0007
98 IPI00009276 PROCR Endothelial protein C receptor precursor 31 kDa 128 135 21 25 14.5 0.0005
99 IPI00027462 S100A9 Protein S100-A9 13 kDa 132 129 55 44 14.6 0.0005
100 IPI00003919 QPCT Isoform 1 of Glutaminyl-peptide 41 kDa 142 132 18 23 14.7 0.0004
cyclotransferase
101 IPI00153049 MXRA8 Isoform 2 of Matrix-remodeling-associated 50 kDa 120 122 17 20 15.0 0.0004
protein 8
102 IPI00297284 IGFBP2 insulin-like growth factor-binding protein 35 kDa 104 93 21 20 15.3 0.0003
2 precursor
103 IPI00647704 IGHA1 cDNA FLJ41552 fis, clone COLON2004478, 53 kDa 27 21 8 8 15.9 0.0003
highly similar to Protein
Troalpha1 H,myeloma
104 IPI00296777 SPARCL1 SPARC-like protein 1 75 kDa 121 116 8 14 15.9 0.0003
105 IPI00304808 KLK1 Isoform 1 of Kallikrein-1 29 kDa 112 108 8 8 16.3 0.0003
106 IPI00796919 GLB1 Beta-galactosidase 82 kDa 126 125 13 11 16.5 0.0003
107 IPI00641737  Haptoglobin 47 kDa 184 176 38 42 16.9 0.0003
108 IPI00152418 CD55 Decay-accelerating factor splicing variant 4 56 kDa 209 228 57 66 17.4 0.0003
109 IPI00395488 VASN Vasorin 72 kDa 165 164 23 24 17.8 0.0003
110 IPI00456429 UBA52 Ubiquitin-60S ribosomal protein L40 15 kDa 192 216 80 88 17.9 0.0003
111 IPI00027310 MEGF8 Isoform 1 of Multiple epidermal growth 303 kDa 147 154 20 15 18.3 0.0003
factor-like domains protein 8
112 IPI00170635 SECTM1 Secreted and transmembrane protein 1 27 kDa 250 253 108 117 18.5 0.0003
113 IPI00294713 MASP2 Isoform 1 of Mannan-binding lectin serine 76 kDa 221 223 67 67 18.7 0.0003
protease 2
114 IPI00007047 S100A8 Protein S100-A8 11 kDa 215 229 87 99 18.8 0.0003
115 IPI00022608 SORL1 Sortilin-related receptor 248 kDa 104 98 4 3 18.9 0.0003
116 IPI00029275 MFI2 Isoform 1 of Melanotransferrin 80 kDa 118 113 2 2 19.0 0.0003
117 IPI00291866 SERPING1 Plasma protease C1 inhibitor 55 kDa 389 430 155 154 19.1 0.0003
118 IPI00009950 LMAN2 Vesicular integral-membrane protein VIP36 40 kDa 248 252 99 108 19.3 0.0003
119 IPI00292150 LTBP2 Latent-transforming growth factor beta- 195 kDa 109 96 4 10 19.5 0.0003
binding protein 2
120 IPI00221224 ANPEP Aminopeptidase N 110 kDa 337 340 122 112 20.7 0.0002
121 IPI00023673 LGALS3BP Galectin-3-binding protein 65 kDa 364 336 121 115 20.9 0.0002
122 IPI00022204 SERPINB3 Isoform 1 of Serpin B3 45 kDa 196 187 24 24 21.0 0.0002
123 IPI00852735 FAT4 Isoform 1 of Protocadherin Fat 4 543 kDa 87 90 0 0 21.2 0.0002
124 IPI00412407 SERPINB4 Serpin peptidase inhibitor, clade B 42 kDa 27 24 0 0 21.2 0.0002
(Ovalbumin), member 4, isoform CRA_a
125 IPI00103871 ROBO4 Isoform 1 of Roundabout homolog 4 107 kDa 191 183 33 23 21.5 0.0002
126 IPI00220813 EFEMP1 Isoform 2 of EGF-containing fibulin-like 54 kDa 306 292 85 83 22.0 0.0002
extracellular matrix protein 1
127 IPI00644296 TNXB Tenascin XB 456 kDa 96 97 0 0 22.8 0.0002
128 IPI00013945 UMOD Isoform 1 of Uromodulin 70 kDa 990 1025 566 573 23.0 0.0002
129 IPI00180707 FREM2 Isoform 1 of FRAS1-related extracellular 351 kDa 121 127 0 3 24.1 0.0002
matrix protein 2
130 IPI00296180 PLAU Isoform 1 of Urokinase-type plasminogen 49 kDa 141 139 4 5 24.2 0.0002
activator
131 IPI00014048 RNASE1 Ribonuclease pancreatic 18 kDa 436 386 136 135 26.7 0.0000
132 IPI00945229 MGAM Maltase-glucoamylase 312 kDa 361 347 79 68 27.7 0.0000
(continued)

Lupus
 Urinary VDBP in lupus nephritis
DJ Go et al.
8
Table 1 Continued
Spectral count

SLE without LN SLE with LN PLGEM

Accession First Second First Second

No number Gene symbols Protein names MW run run run run STN p value

133 IPI00012503 PSAP Isoform Sap-mu-0 of Proactivator 58 kDa 217 207 60 67 28.1 0.0000
polypeptide
134 IPI00291136 COL6A1 Collagen alpha-1(VI) chain 109 kDa 344 314 74 67 28.4 0.0000
135 IPI00016915 IGFBP7 Insulin-like growth factor-binding protein 7 29 kDa 192 193 19 20 28.5 0.0000
136 IPI00004573 PIGR Polymeric immunoglobulin receptor 83 kDa 413 422 89 84 28.9 0.0000
137 IPI00006662 APOD Apolipoprotein D 21 kDa 431 408 169 171 29.1 0.0000
138 IPI00021447 AMY2B Alpha-amylase 2B 58 kDa 35 34 7 6 29.3 0.0000
139 IPI00010675 TFF2 Trefoil factor 2 14 kDa 216 214 26 26 30.1 0.0000
140 IPI00022418 FN1 Isoform 1 of Fibronectin 263 kDa 438 456 132 106 31.1 0.0000
141 IPI00300786  Alpha-amylase 1 58 kDa 31 28 4 6 31.7 0.0000
142 IPI00025476  Pancreatic alpha-amylase 58 kDa 433 427 73 67 31.9 0.0000
143 IPI00328113 FBN1 Fibrillin-1 312 kDa 272 267 40 40 32.3 0.0000
144 IPI00007221 SERPINA5 Plasma serine protease inhibitor 46 kDa 327 323 22 27 33.7 0.0000
145 IPI00011302 CD59 CD59 glycoprotein 14 kDa 420 414 116 110 34.1 0.0000
146 IPI00024284 HSPG2 Basement membrane-specific heparan sulfate 469 kDa 720 744 197 207 46.5 0.0000
proteoglycan core protein
147 IPI00019449 RNASE2 Non-secretory ribonuclease 18 kDa 476 458 53 50 48.2 0.0000
148 IPI00921849  cDNA FLJ57046, highly similar to 72 kDa 23 21 0 0 49.1 0.0000
Lysosomal alpha-glucosidase
149 IPI00293088 GAA Lysosomal alpha-glucosidase 105 kDa 550 519 77 80 51.6 0.0000
150 IPI00215894 KNG1 Isoform LMW of Kininogen-1 48 kDa 1049 1101 279 262 55.8 0.0000
151 IPI00967121 EGF pro-epidermal growth factor isoform 2 130 kDa 707 672 71 65 59.2 0.0000
preproprotein
152 IPI00160130 CUBN Cubilin 399 kDa 940 973 149 137 65.9 0.0000
153 IPI00024292 LRP2 Low-density lipoprotein receptor-related 522 kDa 988 994 165 149 69.0 0.0000
protein 2

cDNA: complementary DNA; ELISA: enzyme-linked immunosorbent assay; Ig: immunoglobulin; LN: lupus nephritis; LMW: low molecular
weight; MW: molecular weight; PLGEM: power law global error model; SLE: systemic lupus erythematosus.
This proteomic analysis was performed in urine samples from eight individuals (four SLE with LN and four SLE without LN). Retinol-binding
protein 4 (RBP4, No 3), transthyretin (TTR, No 16), and vitamin D-binding protein (VDBP, No 17) were selected for further quantitative
measurement by ELISA. Prostaglandin D synthase (PTGDS) was selected as a reference biomarkers because it was previously reported as a
representative biomarker for LN.15,27

Association between urinary biomarkers and (n ¼ 46) were consistent with the aforementioned
laboratory findings in LN findings (Supplementary file 3: Figure S2).
The concentration of each candidate biomarker
was analyzed to correlate disease activity of SLE Prediction of remission of proteinuria in active LN
patients with LN (n ¼ 62) (Figure 2). The bio-
Having demonstrated the performance of the bio-
marker levels increased significantly in direct pro-
marker candidates in differentiating LN status and
portion to the severity of proteinuria, as estimated
reflecting disease activity, we sought to assess the
by UPCR in a random urine sample and renal
predictability of remission in active LN patients
SLEDAI (range 0–16). In particular, VDBP
(n ¼ 21) treated with immunosuppressive therapy.
was best correlated with UPCR (Spearman’s
Seven patients were treated with CYC pulse ther-
rho ¼ 0.661, p < 0.001) and renal SLEDAI
apy, one patient was treated with mycophenolate
(Spearman’s rho ¼ 0.520, p < 0.001). RBP4
mofetil (MMF), and four patients were treated
showed an inverse correlation with eGFR
with tacrolimus. All patients received CS and 13
(Spearman’s rho ¼ 0.500, p < 0.001), but other
patients received ACE inhibitor or ARB. One
biomarker candidates (VDBP, TTR, and PTGDS)
patient was lost to follow-up.
did not. Urine concentrations of any biomarkers
Urine concentrations of RBP4 were lower in
were not associated with serum anti-dsDNA titer
patients who achieved remission within 12 months
and complement levels. The results from the sub- (n ¼ 9) than in those who did not (n ¼ 11) (223.78
group analysis of biopsy-proven LN patients
(98.17–514.01) vs. 3068.81 (1086.80–18210.22) ng/
Lupus
 Urinary VDBP in lupus nephritis
DJ Go et al.
9

Table 2 Baseline characteristics of SLE patients according to LN status in the validation analysis
SLE (n ¼ 121) SLE with LN (n ¼ 62) p value
p value (inactive
SLE without LN SLE with LN (SLE without Inactive LN Active LN LN vs.
(n ¼ 59) (n ¼ 62) LN vs. with LN) (n ¼ 41) (n ¼ 21) active LN)

Age (years) 44.1  12.8 39.7  12.1 0.055 42.0  12.1 35.2  11.2 0.036
Sex (male) 2 (3.4%) 12 (19.4%) 0.009 5 (12.2%) 7 (33.3%) 0.086
SLE disease duration (years) 10.7 (2.9–16.9) 10.6 (4.5–18.4) 0.524 15.4 (6.6–19.4) 6.7 (1.4–14.8) 0.010
LN disease duration (years) – 7.2 (2.1–17.1) – 10.2 (4.6–17.5) 1.5 (0.1–14.0) 0.003
Biopsy-proven LN – 46 (74.2%) – 29 (70.7%) 17 (81.0%) 0.542
WHO classification II – 3 (6.5%) – 3 (10.3%) 0 (0.0%) 0.286
WHO classification III – 4 (8.7%) – 4 (13.8%) 0 (0.0%) 0.281
WHO classification IV – 29 (63.0%) – 16 (55.2%) 13 (76.5%) 0.210
WHO classification V – 10 (21.7%) – 6 (20.7%) 4 (23.5%) 1.000
Hypertension 16 (27.1%) 21 (33.9%) 0.438 12 (29.3%) 9 (42.9%) 0.396
Diabetes mellitus 5 (8.5%) 4 (6.5%) 0.739 3 (7.3%) 1 (4.8%) 1.000
eGFR (ml/min/1.73 m2) 92.8  21.7 77.7  37.0 0.018 76.5  30.0 79.7  47.6 0.795
UPCR – 0.7 (0.1–2.6) – 0.1 (0.1–0.2) 2.4 (0.9–3.6) <0.001
Anti-dsDNA (IU/ml) (0–10) 8.9 (4.2–18.1) 8.9 (4.9–26.2) 0.624 8.0 (4.7–15.4) 17.9 (7.5–49.8) 0.041
C3 (mg/dl) (90.0–180) 81.9  20.1 80.0  25.0 0.661 84.6  25.5 70.9  21.9 0.043
C4 (mg/dl) (10.0–40.0) 16.6  7.9 17.4  9.2 0.617 17.6  8.3 17.1  10.9 0.829
ESR (mm/hour) (0–15) 15.0 (8.0–26.0) 18.0 (10.5–36.0) 0.186 15.0 (10.5–37.0) 22.0 (8.8–34.3) 0.872
Renal SLEDAI – 4.0 (0.0–4.0) – 0.0 (0.0–4.0) 10.0 (4.0–12.0) <0.001
SLEDAI 2.0 (0.0–4.0) 4.0 (2.0–8.0) 0.006 2.0 (0.0-4.0) 12.0 (5.0–14.5) <0.001
Steroid dose (prednisolone 5.00 (2.5–7.5) 7.5 (5.0–12.5) <0.001 5.0 (3.8–10.0) 12.5 (10.0–20.0) <0.001
equivalent, mg/day)
Calcium with vitamin D supplement 9 (15.3%) 21 (33.9%) 0.021 13 (31.7%) 8 (38.1%) 0.777
Bisphosphonate 4 (6.8%) 5 (8.1%) 1.000 4 (9.8%) 1 (4.8%) 0.654
ACE inhibitor or ARB 7 (11.9%) 21 (33.9%) 0.005 8 (19.5%) 13 (61.9%) 0.002
Previous treatment
Cyclophosphamide 5 (8.5%) 26 (41.9%) <0.001 19 (46.3%) 7 (33.3%) 0.418
Mycophenolate mofetil 0 (0.0%) 8 (12.9%) 0.006 3 (7.3%) 5 (23.8%) 0.107
Cyclosporine/tacrolimus 2 (3.4%) 7 (11.3%) 0.164 3 (7.3%) 4 (19.0%) 0.214

ACE: angiotensin-converting enzyme; anti-dsDNA: anti-double-stranded DNA; ARB: angiotensin receptor blocker; C3: complement component
3; C4: complement component 4; ESR: erythrocyte sedimentation rate; eGFR: estimated glomerular filtration rate; LN: lupus nephritis; Scr: serum
creatinine; SLE: systemic lupus erythematosus; SLEDAI: Systemic Lupus Erythematosus Disease Activity Index; UPCR: urine protein creatinine
ratio; WHO: World Health Organization.
Baseline was the time of urine sampling for quantitative measurement of urine biomarkers. eGFR was assessed using the Modification of Diet in
Renal Disease formula, GFR (ml/min/1.73 m2) ¼ 175  (Scr)1.154  (Age)0.203  (0.742 if female)  (1.212 if African American). In patients with
LN, renal biopsy was performed when indicated based on the clinician’s decision. UPCR and renal SLEDAI were not routinely evaluated in SLE
patients without LN. Data are expressed as the median (25% percentile-75% percentile), mean  standard deviation (SD), or n (%). The reference
value of anti-dsDNA antibody, C3, C4 and ESR was shown. p < 0.05 was considered statistically significant.

mgCr, p ¼ 0.020). Other biomarker candidates
(VDBP, TTR, and PTGDS) were not significantly
different between the two groups (Supplementary
file 4: Figure S3). ROC curve showed that urine
concentration of RBP4 had a good predictive
value with an AUC of 0.808. A cutoff value for
RBP4 of 800 ng/mgCr was chosen based on opti-
mization of sensitivity and specificity (sensitivity
0.82 and specificity 0.89). In univariate logistic
regression analysis, elevated urine concentration of
RBP4 was associated with remission failure of pro-
teinuria (odds ratio 0.028, 95% confidence interval
(CI) 0.002 to 0.367, p ¼ 0.007). However, it was not
significant after adjusting for the covariates of age,
sex, disease duration of LN, comorbidities, anti-
dsDNA titer, serum complements, and induction
treatment regimens (Supplementary file 5: Table S2).
Prediction for future proteinuric flare in SLE without
proteinuria
We also assessed the predictability of flare in
patients with inactive LN (n ¼ 41) and without
LN (n ¼ 59). These 100 patients were followed up
for a mean duration of 43.1  6.9 months. In
patients who developed proteinuric flare (n ¼ 15),
mean duration until flare was 23.7  12.7 months.
Twenty-four patients received additional immuno-
suppressive treatments with CS: CYC in three,
MMF in three, azathioprine (AZA) in six, and

Lupus
 Urinary VDBP in lupus nephritis
DJ Go et al.
10

Figure 1 Levels of urinary biomarker candidates according to LN status in SLE patients. Urine levels of all proteins, VDBP (a),
TTR (b), RBP4 (c), and PTGDS (d), were significantly higher in SLE patients with active LN (n ¼ 21) than in those with inactive
LN (n ¼ 41) and without LN (n ¼ 59). Urinary TTR (b), RBP4 (c), and PTGDS (d) in SLE with inactive LN were significantly
elevated compared to those in SLE without LN. Active LN was defined as renal SLEDAI 4 with UPCR  0.5 mg/mg or 24-hour
urine protein  500 mg. Remaining SLE patients were divided according to history of previous renal involvement. Urine concen-
tration of each protein was normalized to urine creatinine concentrations. Graphs represent median levels with the interquartile
range (25th and 75th percentile). Note that the y axis is on the log scale because of the skewness.
LN: lupus nephritis; PTGDS: prostaglandin D synthase; RBP4: retinol-binding protein 4; SLE: systemic lupus erythematosus;
SLEDAI: Systemic Lupus Erythematosus Disease Activity Index; TTR: transthyretin; UPCR: urine protein-creatinine ratio;
VDBP: vitamin D-binding protein.

cyclosporine A (CsA) or tacrolimus in 14 patients. (343.99–2486.83) vs. 414.27 (181.71–803.61)

ACE inhibitor or ARB were given to
15 patients and hydroxychloroquine was given to
75 patients.
SLE patients with flare had higher urine
concentrations of VDBP (4.32 (2.02–28.34) vs.
1.14 (0.23–9.94) ng/mgCr, p ¼ 0.024), RBP4
(141.09 (68.25–333.21) vs. 68.37 (20.19–172.39)
ng/mgCr, p ¼ 0.036), and PTGDS (795.46
Lupus
ng/mgCr, p ¼ 0.008) than those without flare.
ROC analysis using Youden index determined
that the optimal cutoff values for predicting protei-
nuric flare were 1.90 ng/mgCr for VDBP (AUC
0.683, sensitivity 0.80, and specificity 0.60),
51.85 ng/mgCr for RBP4 (AUC 0.671, sensitivity
0.93, and specificity 0.42) and 294.20 ng/mgCr for
PTGDS (AUC 0.715, sensitivity 0.93, and
 Urinary VDBP in lupus nephritis
DJ Go et al.
11

Figure 2 Correlations between urinary biomarker candidates and laboratory measures in SLE patients with LN (n ¼ 62). All
urinary biomarkers (VDBP, TTR, RBP4, and PTGDS) were correlated closely with disease activities of LN, such as UPCR and
rSLEDAI. RBP4 was inversely correlated with eGFR. None of the urinary biomarkers were related to anti-dsDNA titer, serum
complements, and ESR.
anti-dsDNA: anti-double-stranded DNA antibody titer; C3: complement component 3; C4: complement component 4; eGFR:
estimated glomerular filtration rate; ESR: erythrocyte sedimentation rate; LN: lupus nephritis; PTGDS: prostaglandin D synthase;
RBP4: retinol-binding protein 4; rSLEDAI: renal Systemic Lupus Erythematosus Disease Activity Index; SLE: systemic lupus
erythematosus; SLEDAI: Systemic Lupus Erythematosus Disease Activity Index; TTR: transthyretin; UPCR: urine protein/cre-
atinine ratio; VDBP: vitamin D-binding protein.

specificity 0.40) (Figure 3). Using these cutoff
values in a Kaplan-Meier analysis, patients with
higher urine concentrations of biomarker candi-
dates showed a considerably increased rate of
flare compared with those with lower concentra-
tions (Supplementary file 6: Figure S4). In multi-
variate Cox regression, higher urine VDBP (hazard
ratio 9.627, 95% CI 1.698 to 54.571, p ¼ 0.011) was
significantly associated with future development of
significant proteinuria (UPCR  1.0 mg/mg) after
adjustment for age, sex, disease duration of SLE,
previous history of active LN, comorbidities, anti-
dsDNA titer, serum complements, and immuno-
suppressive treatment. Younger age was also at
enhanced risk of developing proteinuria in SLE
(Table 3).
Discussion
The present study has identified urinary biomarker
candidates for LN using proteomics and verified
their clinical utility in SLE patients by ELISA-
based quantification. When analyzed with quanti-
tative measurement in a larger subset of 121 SLE
patients, increased levels of these biomarker candi-
dates could indicate the presence of active LN.
Considering the robust correlation with UPCR or
renal SLEDAI, these proteins are implicated as
organ-specific biomarkers of LN. Increased urine
VDBP was identified as a risk factor predicting
proteinuric flare in SLE patients.
Of the four biomarkers in our panel, VDBP,
TTR, and RBP4 proteins were markedly elevated
in the pooled urinary proteome analysis of SLE
patients with LN compared to those without LN.
These proteins have been reported as potential
predictors of SLE and glomerular disease in the
literature.30,31 However, there were only a few
reports that showed their value in LN.32,33 Until
recently, urine VDBP and TTR have not been
reported to be clinically significant in LN.
Although PTGDS is known as a urinary bio-
marker that is elevated in the urine of patients
Lupus
 Urinary VDBP in lupus nephritis
DJ Go et al.
12

Figure 3 Levels of urinary biomarkers and ROC curves in SLE patients without proteinuria, according to future flare of pro-
teinuria. Among SLE patients (n ¼ 100) without definite proteinuria at the time of urine sampling, including those with inactive LN
(n ¼ 41), proteinuria with UPCR > 1.0 mg/mg occurred in 15 patients. Urine levels of VDBP (a), RBP4 (c), and PTGDS (d) were
significantly higher in patients with future proteinuric flare, but TTR levels (b) were not. ROC curve analysis of VDBP (e), RBP4
(f), and PTGDS (g) was performed to determine optimal cutoff values for prediction of future proteinuric flare. Cutoff values of
urinary biomarkers were 1.89 ng/mgCr in VDBP, 51.85 ng/mgCr in RBP4, and 294.20 ng/mgCr in PTGDS. Graphs represent
median levels with the interquartile range (25th and 75th percentiles). Note that the y axis is on the log scale because of the
skewness. Optimal cutoff value of ROC curve, represented as a black square, was determined at the point for which (sensitivity plus
specificity) is maximal according to Youden-index. Considering characteristics as the screening assay of urinary biomarker, higher
sensitivity was favored for selection of cutoff value.
AUC: area under the curve; LN: lupus nephritis; PTGDS: prostaglandin D synthase; RBP4: retinol-binding protein 4; ROC:
receiver operating characteristic; SLE: systemic lupus erythematosus; TTR: transthyretin; UPCR: urine protein/creatinine ratio;
VDBP: vitamin D-binding protein.

with active LN,27 we unexpectedly detected
decreased PTGDS excretion in the pooled urine
samples of SLE patients with LN in a proteome
analysis. However, the ELISA-based PTGDS
measurement was consistent with other
Lupus
studies.14,15 One possible explanation for the
contradictory quantification readouts between
the two methods could be that the antibody bind-
ing epitope or PTGDS protein isoform recognized
by PTGDS antibody in the ELISA assay is not
 Urinary VDBP in lupus nephritis
DJ Go et al.
13

Table 3 Urinary biomarkers and clinical characteristics predicting future development of proteinuric flare (UPCR > 1.0) in SLE
patients without proteinuria (UPCR < 0.5)
Univariate Cox regression analysis

B Wald HR (95% CI) p value

Sex (male) 1.626 6.172 5.083 (1.409–18.331) 0.013

Age (<40 years) 2.449 10.375 11.576 (2.608–51.371) 0.001
Previous history of active LN 1.846 8.175 6.337 (1.787–22.471) 0.004
Urine VDBP (>1.90 ng/mgCr) 1.677 6.738 5.348 (1.508–18.965) 0.009
Urine RBP4 (>51.85 ng/mgCr) 2.256 4.748 9.543 (1.255–72.593) 0.029
Urine PTGDS (>294.20 ng/mgCr) 2.133 4.245 8.439 (1.109–64.200) 0.039

Multivariate Cox regression analysis

B Wald HR (95% CI) p value

Age (<40 years) 2.809 9.538 16.587 (2.790–98.590) 0.002

Urine VDBP (>1.90 ng/mgCr) 2.265 6.545 9.627 (1.698–54.571) 0.011

anti-dsDNA: anti-double-stranded DNA; CI: confidence interval; eGFR: estimated glomerular filtration rate; HR: hazard ratio; LN: lupus
nephritis; PTGDS: prostaglandin D synthase; RBP4: retinol binding protein 4; SLE: systemic lupus erythematosus; UPCR: urine protein creatinine
ratio; VDBP: vitamin D binding protein.
In multivariate analysis, adjusted covariates were age, sex, previous history of active LN, disease duration of SLE, hypertension, diabetes, renal
insufficiency with eGFR < 60 ml/min/1.73 m2, serologic biomarkers (anti-dsDNA titer and serum complements), additional immunosuppressive
treatments (cyclophosphamide, mycophenolate mofetil, cyclosporine, tacrolimus, or azathioprine), use of angiotensin-converting enzyme inhibitors
or angiotensin receptor blockers, and hydroxychloroquine treatment. Cutoff values of urine biomarkers were obtained from receiver operating
characteristic curve analyses. Variables presenting p value < 0.05 in Cox regression analysis are shown in the table.

necessarily identical with the proteolytic peptide
readout in an MS-based proteome analysis.34
Previously, the clinical utility of elevated urine
VDBP level was mainly investigated in diabetic
nephropathy and non-lupus glomerulonephropa-
thies.35–37 In the present study, urinary VDBP
was closely correlated with disease severity in LN.
When LN was allocated into active LN and inac-
tive LN, the correlation between urine VDBP and
UPCR was significant in active LN patients with
proteinuria (Spearman’s rho ¼ 0.623, p ¼ 0.004) but
not in patients with inactive LN (Spearman’s
rho ¼ 0.273, p ¼ 0.152). In addition, urine VDBP
was predictive of the future development of pro-
teinuria (UPCR  1.0 mg/mg) in SLE patients with-
out proteinuria. The cutoff value (1.90 ng/mgCr) by
ROC analysis was much lower than median VDBP
value (4.32 ng/mgCr) in 15 SLE patients with pro-
teinuric flare. Nevertheless, it has a relatively high
sensitivity (80.0%) and good negative predictive
value (94.4%). This seems to be rather effective in
screening SLE patients who are less likely to
develop flare in the future, when urine VDBP is
below the cutoff value. It may help clinicians to
determine the intensity of immuno-suppressive ther-
apy in these patients. Meanwhile, urine VDBP was
irrelevant as to predictability for the future remis-
sion of active LN in our study. But, as the possi-
bility that urine VDBP level may be a predictor for
achieving remission of LN has been newly raised,38
additional reevaluation with a large number of
patients is required.
VDBP, a multifunctional protein synthesized in
the liver, plays a role in binding and transporting
vitamin D metabolites to target tissues through the
circulation.39 Normally, VDBP is nearly undetect-
able in urine. Glomerular filtration and subsequent
proximal tubular reabsorption of VDBP are critical
processes in the conversion of 25-hydroxy
vitamin D (25(OH)D) into the active metabolite,
1,25-dihydroxyvitamin D (1,25(OH)2D), via mega-
lin/cubilin-mediated endocytosis.40 In patients with
chronic kidney disease, increase of urine VDBP is
associated with tubulointerstitial inflammation,
independently of albuminuria.37 Thus, we suggest
that urinary loss of VDBP in SLE patients without
proteinuria reflects glomerular capillary leakage
and subclinical tubular damage, which reduce
reabsorption of VDBP and diminish 1,25(OH)2D
synthesis. This might stimulate renal macrophages
and promote compensatory production of VDBP,
which can augment the complement-mediated
chemotaxis of neutrophils.41,42 However, we did
not check the markers of tubular damage that
could affect this urine biomarker level. Although
not evaluated in our study, it was reported that
serum 25(OH)D levels were inversely associated
with urine VDBP concentrations in patients with
Lupus
 Urinary VDBP in lupus nephritis
DJ Go et al.
14
pediatric SLE.30 In another study, SLE patients
with active LN had significantly lower serum
levels of 25(OH)D with an inverse correlation
with UPCR.43 In vitro, 1,25(OH)2D has a regula-
tory effect on B-lymphocyte activation and auto-
antibody production that is closely involved in the
pathogenesis of LN.44 These findings suggest that
urine VDBP could be useful for the forecast of
renal flare in SLE patients. However, whether urin-
ary loss of VDBP actually affects vitamin D status
is debatable.35,45 Further evaluation of serum
VDBP, 25(OH)D and 1,25(OH)2D in patients
with SLE may be required to demonstrate the prog-
nostic value of VDBP.
Urine concentrations of TTR, RBP4, and
PTGDS proteins remained higher in patients with
inactive LN compared to SLE patients without LN
(Figure 1). These proteins are assumed to be ele-
vated because of remaining subclinical kidney
injury after clinical improvement of active LN.
Thus, they may be more sensitive to nephritis
than UPCR. In the previous study, urinary excre-
tion of TTR, a carrier protein of thyroxine
and RBP in blood,46 predicted the response to
ACE inhibitor therapy in immunoglobulin A
nephropathy.47 In LN, TTR was associated with
severity of concurrent proteinuria, but not with
prognosis of future proteinuria. Urinary RBP was
suggested as an independent prognostic marker
that could identify patients who will progress to
renal insufficiency in non-lupus glomerulonephro-
pathies.48 Also, it is reported to be a determining
factor of steroid responsiveness in patients with
idiopathic nephrotic syndrome.49 In LN, urine
RBP is reportedly a sensitive marker of proximal
tubular dysfunction.32 In the present study, urine
RBP4 level inversely correlated with eGFR and
was associated with risk of remission failure in
patients with active LN in our univariate regression
analysis. However, it was not a significant factor
after adjusting for baseline characteristics and
concurrent immunosuppressive treatments. The
study was not powered for statistical conclu-
siveness, being based on only 20 patients with
active LN.
We included a relatively large cross-sectional
cohort (n ¼ 121) of SLE patients in validation with
quantitative measurement and a four-year follow-
up. Urine biomarkers were superior to conventional
serologic markers, such as anti-dsDNA and comple-
ment, for evaluation of disease activity of LN. Their
diagnostic values were significant both in LN
according to the ACR criteria and biopsy-proven
LN. To our knowledge this study is the first to
Lupus
investigate urine VDBP as a long-term prognostic
biomarker for renal involvement of SLE.
There are some limitations of our study. In a
screening study, as patients with LN received
more doses of CS at time of urine sampling, the
possibility that immunosuppressive therapy
affected the proteomics result cannot be excluded.
But, there was no difference according to the cur-
rent treatment in patients with LN in a validation
study. Baseline characteristics were not identical
within groups in a validation study, even though
we adjusted them in regression analysis. There
was a lack of sufficient patient numbers with
future development of proteinuria (n ¼ 15) while
there were a number of parameters for multivariate
analysis. Assessment of histologic findings of LN
was precluded because urine sampling was not per-
formed simultaneously with renal biopsy. In some
LN patients (25.8%), LN was not confirmed by
renal biopsy because of the invasiveness of the pro-
cedure. It was not possible to trace change in urine
biomarkers as a response to treatment because lon-
gitudinal measurement was not conducted.
Urinary biomarkers including VDBP reflect LN
status and renal disease activity in SLE patients.
Urine VDBP concentrations may allow clinicians
to anticipate the occurrence of future renal flare
in SLE patients with clinically quiescent renal
involvement. To confirm these findings, prospective
controlled studies are needed.
Declaration of conflicting interests
The authors declared no potential conflicts of inter-
est with respect to the research, authorship, and/or
publication of this article.
Funding
The author(s) disclosed receipt of the following
financial support for the research, authorship,
and/or publication of this article: This study was
supported by a grant of the Korea Health
Technology R&D Project through the Korea
Health Industry Development Institute (KHIDI),
funded by the Ministry of Health & Welfare,
Republic of Korea (grant number : HI14C1277)
and grants from the Ministry of Science, ICT and
Future Planning, Republic of Korea (NRF-
2015M3A9B6052011, 2015M3A9B6073835 &
2016R1A5A1010764).

ORCID iD
DJ Go http://orcid.org/0000-0001-7350-3491
References
1 Ortega LM, Schultz DR, Lenz O, Pardo V, Contreras GN. Review:
Lupus nephritis: Pathologic features, epidemiology and a guide to
therapeutic decisions. Lupus 2010; 19: 557–574.
2 Korbet SM, Lewis EJ, Schwartz MM, Reichlin M, Evans J, Rohde
RD. Factors predictive of outcome in severe lupus nephritis. Lupus
Nephritis Collaborative Study Group. Am J Kidney Dis 2000; 35:
904–914.
3 Hill GS, Delahousse M, Nochy D, et al. Outcome of
relapse in lupus nephritis: Roles of reversal of renal fibrosis and
response of inflammation to therapy. Kidney Int 2002; 61:
2176–2186.
4 Fiehn C, Haijar Y, Mueller K, Waldherr R, Ho AD, Andrassy K.
Improved clinical outcome of lupus nephritis during the past
decade: Importance of early diagnosis and treatment. Ann Rheum
Dis 2003; 62: 435–439.
5 Faurschou M, Starklint H, Halberg P, Jacobsen S. Prognostic
factors in lupus nephritis: Diagnostic and therapeutic delay
increases the risk of terminal renal failure. J Rheumatol 2006; 33:
1563–1569.
6 Stiles KP, Yuan CM, Chung EM, Lyon RD, Lane JD, Abbott KC.
Renal biopsy in high-risk patients with medical diseases of the
kidney. Am J Kidney Dis 2000; 36: 419–433.
7 Eiro M, Katoh T, Watanabe T. Risk factors for bleeding compli-
cations in percutaneous renal biopsy. Clin Exp Nephrol 2005; 9:
40–45.
8 Moroni G, Radice A, Giammarresi G, et al. Are laboratory tests
useful for monitoring the activity of lupus nephritis? A 6-year pro-
spective study in a cohort of 228 patients with lupus nephritis. Ann
Rheum Dis 2009; 68: 234–237.
9 Reyes-Thomas J, Blanco I, Putterman C. Urinary biomarkers in
lupus nephritis. Clin Rev Allergy Immunol 2011; 40: 138–150.
10 Rubinstein T, Pitashny M, Levine B, et al. Urinary neutrophil
gelatinase-associated lipocalin as a novel biomarker for disease
activity in lupus nephritis. Rheumatology (Oxford) 2010; 49:
960–971.
11 Torres-Salido MT, Cortés-Hernández J, Vidal X, Pedrosa A,
Vilardell-Tarrés M, Ordi-Ros J. Neutrophil gelatinase-associated
lipocalin as a biomarker for lupus nephritis. Nephrol Dial
Transplant 2014; 29: 1740–1749.
12 Rosa RF, Takei K, Araújo NC, Loduca SM, Szajubok JC,
Chahade WH. Monocyte chemoattractant-1 as a urinary bio-
marker for the diagnosis of activity of lupus nephritis in
Brazilian patients. J Rheumatol 2012; 39: 1948–1954.
13 Schwartz N, Rubinstein T, Burkly LC, et al. Urinary TWEAK as a
biomarker of lupus nephritis: A multicenter cohort study. Arthritis
Res Ther 2009; 11: R143.
14 Somparn P, Hirankarn N, Leelahavanichkul A, Khovidhunkit W,
Thongboonkerd V, Avihingsanon Y. Urinary proteomics revealed
prostaglandin H(2)D-isomerase, not Zn-a2-glycoprotein, as a bio-
marker for active lupus nephritis. J Proteomics 2012; 75:
3240–3247.
15 Gupta R, Yadav A, Misra R, Aggarwal A. Urinary prostaglandin
D synthase as biomarker in lupus nephritis: A longitudinal study.
Clin Exp Rheumatol 2015; 33: 694–698.
16 Korte EA, Gaffney PM, Powell DW. Contributions of mass spec-
trometry-based proteomics to defining cellular mechanisms and
diagnostic markers for systemic lupus erythematosus. Arthritis
Res Ther 2012; 14: 204.
17 Hochberg MC. Updating the American College of Rheumatology
revised criteria for the classification of systemic lupus erythemato-
sus. Arthritis Rheum 1997; 40: 1725.
Urinary VDBP in lupus nephritis
DJ Go et al.
15
18 Weening JJ, D’Agati VD, Schwartz MM, et al. The classification
of glomerulonephritis in systemic lupus erythematosus revisited.
J Am Soc Nephrol 2004; 15: 241–250.
19 Gladman DD, Ibañez D, Urowitz MB. Systemic lupus erythe-
matosus disease activity index 2000. J Rheumatol 2002; 29:
288–291.
20 Rovin BH, Furie R, Latinis K, et al. Efficacy and safety of ritux-
imab in patients with active proliferative lupus nephritis: The
Lupus Nephritis Assessment with Rituximab study. Arthritis
Rheum 2012; 64: 1215–1226.
21 Hebert LA, Dillon JJ, Middendorf DF, Lewis EJ, Peter JB.
Relationship between appearance of urinary red blood cell/white
blood cell casts and the onset of renal relapse in systemic lupus
erythematosus. Am J Kidney Dis 1995; 26: 432–438.
22 Gordon C, Jayne D, Pusey C, et al. European consensus statement
on the terminology used in the management of lupus glomerulo-
nephritis. Lupus 2009; 18: 257–263.
23 Afkarian M, Bhasin M, Dillon ST, et al. Optimizing a proteomics
platform for urine biomarker discovery. Mol Cell Proteomics 2010;
9: 2195–2204.
24 Shevchenko A, Tomas H, Havlis J, Olsen JV, Mann M. In-gel
digestion for mass spectrometric characterization of proteins and
proteomes. Nat Protoc 2006; 1: 2856–2860.
25 Nesvizhskii AI, Keller A, Kolker E, Aebersold R. A statistical
model for identifying proteins by tandem mass spectrometry.
Anal Chem 2003; 75: 4646–4658.
26 Pavelka N, Fournier ML, Swanson SK, et al. Statistical similarities
between transcriptomics and quantitative shotgun proteomics
data. Mol Cell Proteomics 2008; 7: 631–644.
27 Wu T, Fu Y, Brekken D, et al. Urine proteome scans uncover total
urinary protease, prostaglandin D synthase, serum amyloid P, and
superoxide dismutase as potential markers of lupus nephritis.
J Immunol 2010; 184: 2183–2193.
28 Suzuki M, Wiers K, Brooks EB, et al. Initial validation of a novel
protein biomarker panel for active pediatric lupus nephritis.
Pediatr Res 2009; 65: 530–536.
29 Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D.
A more accurate method to estimate glomerular filtration rate
from serum creatinine: A new prediction equation. Modification
of Diet in Renal Disease Study Group. Ann Intern Med 1999; 130:
461–470.
30 Robinson AB, Thierry-Palmer M, Gibson KL, Rabinovich CE.
Disease activity, proteinuria, and vitamin D status in children
with systemic lupus erythematosus and juvenile dermatomyositis.
J Pediatr 2012; 160: 297–302.
31 Varghese SA, Powell TB, Budisavljevic MN, et al. Urine bio-
markers predict the cause of glomerular disease. J Am Soc
Nephrol 2007; 18: 913–922.
32 Sesso R, Rettori R, Nishida S, Sato E, Ajzen H, Pereira AB.
Assessment of lupus nephritis activity using urinary retinol-binding
protein. Nephrol Dial Transplant 1994; 9: 367–371.
33 Guy JM, Brammah TB, Holt L, et al. Urinary excretion of albumin
and retinol binding protein in systemic lupus erythematosus. Ann
Clin Biochem 1997; 34: 668–674.
34 Oonk S, Spitali P, Hiller M, et al. Comparative mass spectrometric
and immunoassay-based proteome analysis in serum of Duchenne
muscular dystrophy patients. Proteomics Clin Appl 2016; 10:
290–299.
35 Thrailkill KM, Jo CH, Cockrell GE, Moreau CS, Fowlkes JL.
Enhanced excretion of vitamin D binding protein in type 1 dia-
betes: A role in vitamin D deficiency?. J Clin Endocrinol Metab
2011; 96: 142–149.
36 Tian XQ, Zhao LM, Ge JP, Zhang Y, Xu YC. Elevated urinary
level of vitamin D-binding protein as a novel biomarker for dia-
betic nephropathy. Exp Ther Med 2014; 7: 411–416.
37 Mirković K, Doorenbos CR, Dam WA, et al. Urinary vitamin D
binding protein: A potential novel marker of renal interstitial
inflammation and fibrosis. PLoS One 2013; 8: e55887.
38 Brunner HI, Bennett MR, Gulati G, et al. Urine biomarkers to
predict response to lupus nephritis therapy in children and young
adults. J Rheumatol 2017; 44: 1239–1248.
Lupus
 Urinary VDBP in lupus nephritis
DJ Go et al.
16
39 Verboven C, Rabijns A, De Maeyer M, Van Baelen H, Bouillon R,
De Ranter C. A structural basis for the unique binding features of
the human vitamin D-binding protein. Nat Struct Biol 2002; 9:
131–136.
40 Nykjaer A, Dragun D, Walther D, et al. An endocytic pathway
essential for renal uptake and activation of the steroid 25-(OH)
vitamin D3. Cell 1999; 96: 507–515.
41 Schiffer L, Bethunaickan R, Ramanujam M, et al. Activated renal
macrophages are markers of disease onset and disease remission in
lupus nephritis. J Immunol 2008; 180: 1938–1947.
42 Chishimba L, Thickett DR, Stockley RA, Wood AM. The vitamin
D axis in the lung: A key role for vitamin D-binding protein.
Thorax 2010; 65: 456–462.
43 Sumethkul K, Boonyaratavej S, Kitumnuaypong T, et al. The pre-
dictive factors of low serum 25-hydroxyvitamin D and vitamin D
deficiency in patients with systemic lupus erythematosus.
Rheumatol Int 2013; 33: 1461–1467.
44 Chen S, Sims GP, Chen XX, Gu YY, Chen S, Lipsky PE.
Modulatory effects of 1,25-dihydroxyvitamin D3 on human B
cell differentiation. J Immunol 2007; 179: 1634–1647.
Lupus
45 Doorenbos CR, de Cuba MM, Vogt L, et al. Antiproteinuric treat-
ment reduces urinary loss of vitamin D-binding protein but does
not affect vitamin D status in patients with chronic kidney disease.
J Steroid Biochem Mol Biol 2012; 128: 56–61.
46 Kazemipour N, Qazizadeh H, Sepehrimanesh M, Salimi S.
Biomarkers identified from serum proteomic analysis for the dif-
ferential diagnosis of systemic lupus erythematosus. Lupus 2015;
24: 582–587.
47 Rocchetti MT, Centra M, Papale M, et al. Urine protein profile of
IgA nephropathy patients may predict the response to ACE-inhi-
bitor therapy. Proteomics 2008; 8: 206–216.
48 Kirsztajn GM, Nishida SK, Silva MS, Ajzen H, Moura LA,
Pereira AB. Urinary retinol-binding protein as a prognostic
marker in glomerulopathies. Nephron 2002; 90: 424–431.
49 Sesso R, Santos AP, Nishida SK, Klag MJ, Carvalhaes JT, Ajzen
H. Prediction of steroid responsiveness in the idiopathic nephrotic
syndrome using urinary retinol-binding protein and beta-2-micro-
globulin. Ann Intern Med 1992; 116: 905–909.