A Pilot Trial Assessing Urinary Gene Expression Profiling with an mRNA Array for Diabetic Nephropathy
et al. (2012) A Pilot Trial Assessing Urinary Gene Expression Profiling with an mRNA Array for Diabetic
Nephropathy. PLoS ONE 7(5): e34824. doi:10.1371/journal.pone.0034824
A Pilot Trial Assessing Urinary Gene Expression Profiling with an mRNA Array for Diabetic Nephropathy
Min Zheng 0
Lin-Li Lv 0
Yu-Han Cao 0
Hong Liu 0
Jie Ni 0
Hou-Yong Dai 0
Dan Liu 0
Xiang-Dong Lei 0
Cheng Liu 0
John E. Mendelson, California Pacific Medicial Center Research Institute, United States of America
0 1 Institute of Nephrology, Zhong Da Hospital, Southeast University School of Medicine , Nanjing, China, 2 CT Bioscience, Chang Zhou , China
Background: The initiation and progression of diabetic nephropathy (DN) is complex. Quantification of mRNA expression in urinary sediment has emerged as a novel strategy for studying renal diseases. Considering the numerous molecules involved in DN development, a high-throughput platform with parallel detection of multiple mRNAs is needed. In this study, we constructed a self-assembling mRNA array to analyze urinary mRNAs in DN patients with aims to reveal its potential in searching novel biomarkers. Methods: mRNA array containing 88 genes were fabricated and its performance was evaluated. A pilot study with 9 subjects including 6 DN patients and 3 normal controls were studied with the array. DN patients were assigned into two groups according to their estimate glomerular rate (eGFR): DNI group (eGFR.60 ml/min/1.73 m2, n = 3) and DNII group (eGFR,60 ml/min/1.73 m2, n = 3). Urinary cell pellet was collected from each study participant. Relative abundance of these target mRNAs from urinary pellet was quantified with the array. Results: The array we fabricated displayed high sensitivity and specificity. Moreover, the Cts of Positive PCR Controls in our experiments were 2460.5 which indicated high repeatability of the array. A total of 29 mRNAs were significantly increased in DN patients compared with controls (p,0.05). Among these genes, a-actinin4, CDH2, ACE, FAT1, synaptopodin, COL4a, twist, NOTCH3 mRNA expression were 15-fold higher than those in normal controls. In contrast, urinary TIMP-1 mRNA was significantly decreased in DN patients (p,0.05). It was shown that CTGF, MCP-1, PAI-1, ACE, CDH1, CDH2 mRNA varied significantly among the 3 study groups, and their mRNA levels increased with DN progression (p,0.05). Conclusion: Our pilot study demonstrated that mRNA array might serve as a high-throughput and sensitive tool for detecting mRNA expression in urinary sediment. Thus, this primary study indicated that mRNA array probably could be a useful tool for searching new biomarkers for DN.
. These authors contributed equally to this work.
Diabetic nephropathy (DN) is a serious complication of diabetes
with high morbidity and mortality. It has become the most
common cause of end-stage kidney disease in the world .
Therefore, the early identification of people at great risk of DN is
of the utmost importance.
Initial increase in albumin excretion rate (AER) has been
traditionally linked to a subsequent decline in glomerular filtration
rate (GFR) and microalbuminuria (MA) has been established as an
early biomarker of DN in clinic . However, approximately 20%
of people with type 2 diabetes develop at least stage 3 CKD,
defined as an estimated GFR (eGFR) less than 60 ml/min/
1.73 m2, while remaining normoalbuminuric. This discordance
between changes in GFR and AER has resulted in a search for
new markers that identify people with diabetes who are at risk of
declining GFR independent of progressive increases in AER .
In the past decade, evidence from in vitro experiments and
pathological examinations to epidemiological studies, has shown
that pathogenetic mechanism of DN involved complex molecular
pathways. The key molecules and pathways included
inflammation , epithelial-to-mesenchymal transition , extra-cellular
matrix deposition, tubular cell and podocyte injury [6,7]. Target
screening of those related molecules might provide us an efficient
strategy for novel biomarker discovery.
Recently, RNA extraction from urinary sediment combined
with real-time quantitative PCR has emerged as a novel strategy
for identifying biomarkers with kidney disease. Preliminary studies
have suggested that the determination of urinary mRNA levels
might be valuable in monitoring the progression of renal disease
[8,9]. Our previous studies have also indicated the potential
DNI (N = 3)
DNII (N = 3)
application of urinary mRNA detection in searching novel
biomarkers of kidney disease [10,11]. However, considering the
numerous molecules involved in DN development, a
highthroughput platform with efficient detection capacity is needed.
Gene expression array has emerged as a promising strategy for
analyzing multiple gene expression in parallel . Among the
various array formats, PCR array is the most reliable tools for
analyzing the expression of a focused panel of genes . The
arrays can be run efficiently in parallel, enabling studies on the
large populations of molecules involved in disease development
that are necessary for marker discovery and validation.
In this study, a target PCR array was constructed with the aim
to develop a simple, easy-to-use tool for differential mRNA
screening with urinary sediment. And the PCR array was applied
to profile the expression of a panel of mRNAs relevant to pathways
involved in DN pathophysiology. The screening with such
platform might provide important clues as for candidate gene
markers for further validation. To our best knowledge, this is the
first study to apply PCR arrays to analyze urinary mRNA
expression with aims to screening promising genes for DN
1 Characterization of Real Time PCR Array
1.1 Sensitivty. We assessed sensitivity of PCR array in terms
of positive call rate, which means the percentage of genes
detectable in an array. Figure S1 shown the percentage of positive
calls (the percentage of genes with Ct,38) against different
samples. The results indicate that the PCR array system achieved
greater than 80 percent positive calls with all samples, and 5 out of
9 over 90% positive calls, 8 out of 9 over 85%.
1.2 Specificity. Figure S2 shown the melting curve analysis
which was used to assess the specificity of the array. As displayed in
the figure, a single product of the predicted size from each reaction
without secondary products such as primer dimers indicated the
high specificity of PCR array.
1.3 Reproducibility. The PPC contains artificial DNA
template which has no homology to the DNA of the samples
and a validated primer pair which can be used to check
plate-toplate and run-to-run reproducibility of the array. Both
plate-toplate or run-to-run variation will result in PPC Ct variation. As
shown in Figure S3, the Cts of PPC in our experiments were
2460.5. The results demonstrated that high degree of
plate-toplate and run-to-run reproducibility could be obtained.
2 Clinical Characteristics of Enrolled Subjects
The primary clinical and laboratory characteristics of the study
subjects are summarized in Table 1. The age of participants in the
control group was lower than DN patients. DNII group had
a significant increase in urinary albumin excretion, serum
creatinine, BUN and a decrease in eGFR compared with DNI
group and normal controls (p,0.05).
3 Urinary mRNA Profiles in DN Patients and Healthy Controls
Among the 88 mRNAs screened, a total of 29 mRNAs were
significantly increased in DN patients compared with normal
controls (p,0.05) while TIMP-1 was significantly decreased. The
increased 29 mRNAs belong to different functional categories in
DN development as listed in Table 2. If we divide DN group as
subgroup based on the eGFR level, among those 30 mRNAs
identified, 14 mRNAs (SYNPO, ACTN4, PODXL, SNAI2,
UMOD, MAPK8, RBP4, MMP9, CTGF, MCP-1, PAI-1, ACE,
eGFR (ml/min/1.73 m2) 87.8469.03
#DNII group and DNI group vs. controls, p,0.05.
*DNII group vs. DNI group and controls, p,0.05.
CDH1 and CDH2) were found to show significant difference
when DNI and DNII group was compared with normal controls.
And other 14 (HGF, FAT1, PODXL2, MMP2, COL4A1,
TWIST1, SNAI1, SMAD4, SMAD7, LAMA5, REN, NOTCH2,
NOTCH3 and AGT) mRNAs show significant different level
when DNII group was compared with normal controls but no
difference was found when DNI group was compared with normal
controls. The other 2 mRNAs (TIMP-1, cytokeratin18) show no
difference when neither two subgroups were compared with
4 Urinary Gene Expression between Different Stages of DN
We also evaluated the differential expression of mRNAs
between different stages of DN, that is DNI and DNII group.
Interestingly, 6 mRNAs shown significant difference when DNI
and DNII group was compared. The mRNAs identified included
CTGF (p = 0.01), MCP-1(p = 0.04), PAI-1(p = 0.04), ACE
(p,0.001), CDH1 (p = 0.04) and CDH2 (p = 0.03). Moreover,
those mRNAs were found to be differentially expressed when DNI
and DNII group were compared with healthy controls (p,0.05).
CDH1, CDH2, FAT1, LAMA5, cytokeratin18
PODXL2, SYNPO, PODXL, ACTN4
NOTCH2, NOTCH3, TWIST1, SNAI1, SMAD4,
SMAD7, SNAI2, MAPK8
Extra-cellular matrix related
PAI-1, MMP-2, COL4A1, MMP-9
Functional categories of genes up-regulated in DN patients compared with
healthy controls. Among the 88 mRNAs screened, a total of 29 mRNAs were
significantly increased in DN patients compared with normal controls (p,0.05).
5 Fold Change Expressions of mRNAs between DN Patients and Controls
Figure 1 shown the fold change expressions of mRNAs between
DN patients and controls. In our study, a 15 fold change value was
chosen as a threshold. Those mRNAs with 15 fold increased levels
when DN patients were compared with normal controls included 8
genes as followings: NOTCH3, ACTN4, CDH2, ACE, FAT1,
COL4A1, SYNPO, TWIST1. And TIMP-1 was found with 15
fold decreased levels in DN group compared with normal controls.
No candidate was found with 15 fold change when DNI group was
compared with DNII group. Fourteen mRNAs were found with
15 fold change when DNII group was compared with normal
controls. Four mRNAs shown significant change in expression
levels when DNI was compared with controls (NOTCH3,
ACTN4, CDH2, ACE).
In this study, we established an mRNA profiling array specific
for DN using real-time PCR technique and evaluated its
performance. The array was used to screen urinary mRNA levels
in DN patients. The preliminary results demonstrate its potential
to define new biomarkers for kidney diseases.
Substantial effort has been spent to identify gene biomarkers
that display differential expression in various disease types.
However, gene expression studies in kidney disease are usually
limited to the expression of a few protein-coding genes .
Considering the complex molecular networks involved in kidney
disease, an enormous number of mRNAs may have to be
examined. Array technique with high throughput is a potentially
efficient approach for simultaneously analyzing large-scale gene
expression profiles. Rodder et al found that a transcriptomic
classifier consisting of 19 metzincins and related genes (MARGS)
could discriminate biopsies from renal transplant patients with or
without interstitial fibrosis/tubular atrophy by virtue of gene
expression measurement . In a recent study from Wu et al,
cDNA microarray analysis was performed to identify gene
expression changes, and highly expressed genes were evaluated
as markers both in mice and human kidney samples with
membranous glomerulonephropathy (MN) . However, in
these studies, gene arrays were applied to analyze gene expressions
in renal tissue, which was an invasive assay. Urine containing cells
shed from the urinary tract might provide important information
about ongoing kidney damage [17,18]. In this study, we have
designed a PCR array for urinary gene expression analysis and
conducted a proof-of-concept study with samples from DN
The results demonstrated that the array has relatively high
specificity, sensitivity and reproducibility. The high sensitivity
would suggest that it could quantify mRNA expression with low
amounts of extracted RNA from urinary sediment, including
detection of inflammatory cytokine and receptor mRNAs that
are known to be expressed at very low levels but play a critical
role in DN development. The specificity and reproducibility of
the system suggests the amplification of only one gene-specific
product in each reaction, making the technique suitable for
large-scale gene expression screening. Eighty-eight mRNAs were
included in the constructed array targeting key molecules
thought to be involved in DN development. The upregulated
mRNAs were functionally associated with key molecular events
with the development of DN including those from tubular cell
and podocyte injury markers, renin-angiotensin system, EMT
markers, cytokines, extra-cellular matrix related markers and
signal pathway related mRNAs.
Among the mRNAs we identified, six (CTGF, MCP-1, PAI-1,
ACE, CDH1 and CDH2) showed significant difference when the
DNI and DNII groups were compared, suggesting these mRNAs
might be involved in the development of DN and could hold
promise in discriminating different stages of DN. Many previous
studies have suggested that profibrotic growth factors,
inflammaFigure 1. Fold change expressions of mRNAs between DN patients and controls. NOTCH3, ACTN4, CDH2, ACE, FAT1, COL4A1, SYNPO,
TWIST1 were 15 fold increased and TIMP-1 was found with 15 fold decreased levels in DN group compared with normal controls.
tory chemokines and renin angiotensin system (RAS) are key
players in the pathogenesis of DN . CDH belongs to
cellcell adhesion molecule family and has become a marker for
tracking EMT in development of kidney damage . Hence, our
results are compatible with dysregulation of these genes and
pathways in the pathophysiology of DN. However, our findings
are preliminary and need to be validated in larger cohort group of
In summary, we have constructed a PCR array platform with
high sensitivity, specificity and reproducibility to enable urinary
mRNA profiling. In this proof-of-concept study, we demonstrated
that multiple mRNAs in urine sediment can be profiled in patients
and healthy controls. A few promising mRNAs with differential
expression in DN patients compared with healthy controls were
found suggesting biomarker candidates for further study.
Materials and Methods
Fabrication of PCR Array
88 potential molecules involved in the pathogenesis and
progression of DN were selected as the target mRNAs. GAPDH,
b2-MG, b-actin, RPL27, HPRT1 and OAZ1 were used as house
keeping genes for normalization. A Genomic DNA Control (GDC)
primer was also set to specifically detect non-transcribed, repetitive
genomic DNA. The replicate Positive PCR Controls (PPC) was
included to report on the efficiency of the polymerase chain
reaction itself. The corresponding gene name was listed in part of
Primer design and selection.
Primer Design and Selection
Except for three genes, primers were designed to cover all
transcripts of each gene. Table S1 lists the refseq accession IDs
each primer pair can detect. Primer design also takes the following
into considerations. All primers have similar melting temperature
(Tm), primers do not contain known SNPs or locate in genomic
repetitive regions. Primers are experimentally selected if they meet
the following criterias: a typical amplication curve, post PCR
melting curve analysis shows a single peak, single band with
matched sizes analysed by agarose gel electrophoresis.
All studies were approved by the Ethical Committee of
Southeast University. Written informed consents were obtained
from all subjects to use their urine for research purpose.
Six patients with type 2 diabetic nephropathy from Zhong Da
Hospital, Southeast University School of Medicine were enrolled
in this study. Diabetic nephropathy was diagnosed based on
KDOQI guideline , that is: at least 5 years from the
diagnosis of type 2 diabetes, the presence of diabetic
retinopathy, elevated albumin-creatinine ratio (ACR), and the absence
of clinical or laboratory evidence of other kidney disease. To
evaluate progression of DN, patients were divided into two
groups based on the eGFR level: DNI (eGFR.60, n = 3) and
DNII (eGFR,60, n = 3). eGFR was calculated by the equation
proposed by Ma et al., which was considered to be more
suitable for Chinese study subjects . A group of healthy
volunteers (n = 3) was also enrolled in the study as a negative
control. Clinical data including albuminuria, blood urea
nitrogen (BUN), and serum creatinine were recorded at baseline
for each of the groups.
Collection of Urine Samples and Total RNA Extraction
A whole-stream early-morning urine specimen was collected
from each study participant. Shortly after collection, the urine was
centrifuged at 3,000*g for 30 minutes at 4uC. The urinary
supernatant was discarded, and the remaining cell pellet was
resuspended in 1.5 ml DEPC-treated PBS and was then
centrifuged at 13,000*g for 5 minutes at 4uC. The pellet was then
resuspended in 1.0 ml RNAiso Plus (TAKARA, Dalian, China)
and was stored at 280uC until use. Total RNA was extracted
according to the manufacturers protocol (TAKARA). The RNA
concentration and purity were confirmed using the relative
absorbance ratio at 260/280 on a nanodrop 2000 (Thermo,
For reverse transcription, 2 mg total RNA was mixed with 8 ml
56 PrimeScriptTM Buffer, 2 ml PrimeScriptTM RT Enzyme Mix I,
2 ml Oligo dT Primer (50 mM), 2 ml Random 6 mers (100 mM),
(TAKARA), the solution was increased to a volume of 40 ml with
dH2O. Reverse transcription was performed at 37uC for 15
minutes, followed by an inactivation reaction at 85uC for 5
seconds. The resulting cDNA was stored at 220uC until use.
Each cDNA was diluted using ddH2O to 1000 ul and mixed
with 1 ml 26 SYBR Premix Ex TaqTM (TAKARA, Dalian,
China). An average 20 ml of the mixture was added to each well of
the 96 well PCR array except GDC control. The sealed PCR plate
was loaded in Roche LightCycler 480 II instrument. The cycling
condition is as following: 95uC for 5 min, 45 cycles at 95uC for
10 s, 60uC for 10 s and 72uC for 10 s. Then, dissociation curves
(DC) and melting temperatures (Tm) were recorded. Melting
curve analysis was performed at 95uC for 5 s, 65uC for 1 min and
95uC for 0 sec. Relative changes in gene expression were
calculated using the ggCt (threshold cycle) method. Here,
ggCt = (Ct target-sample - Ct ref-sample ) - (Ct target-control
Ct ref-control). Fold change values were calculated using the
formula as follows: expression fold changes = Target gene
expression level of sample/Target gene expression level of control
SPSS 13.0 was used for data analysis. All results are presented as
mean6SD unless otherwise specified. Baseline data were
compared by a one-way analysis of variance (ANOVA) between
groups. The T-test was used for gene comparison between two
groups. All p-values were two tailed, and a value ,0.05 was
considered to be statistically significant.
Figure S2 Specificity evaluation of PCR array. Melting
curve analysis shown that single peak could be obtained for each
reaction which indicated the high specificity of PCR array.
Figure S3 Reproducibility evaluation of PCR array. The
bar repesents Ct value of PPC for each sample. The Cts of PPC in
our experiments were 2460.5 which demonstrated that high
degree of plate-to-plate and run-to-run reproducibility could be
Conceived and designed the experiments: B-CL MZ. Performed the
experiments: MZ L-LL JN X-DL. Analyzed the data: MZ H-YD.
Contributed reagents/materials/analysis tools: DL HL Y-HC. Wrote the
paper: MZ L-LL B-CL.
Gene Table and Transcripts Detected by the
1. Atkins RC , Zimmet P ( 2010 ) World Kidney Day 2010: diabetic kidney diseaseact now or pay later . Am J Kidney Dis 55 : 205 - 8 .
2. Chang SS ( 2008 ) Albuminuria and diabetic nephropathy . Pediatr Endocrinol Rev 5 Suppl 4 : 974 - 9 .
3. Macisaac RJ , Jerums G ( 2011 ) Diabetic kidney disease with and without albuminuria . Curr Opin Nephrol Hypertens 20 : 246 - 57 .
4. Navarro-Gonzalez JF , Mora-Fernandez C , Muros de Fuentes M , Garcia-Perez J ( 2011 ) Inflammatory molecules and pathways in the pathogenesis of diabetic nephropathy . Nat Rev Nephrol 7 : 327 - 40 .
5. Li MX , Liu BC ( 2007 ) Epithelial to mesenchymal transition in the progression of tubulointerstitial fibrosis . Chin Med J (Engl) 120 : 1925 - 30 .
6. Tesch GH , Lim AK ( 2011 ) Recent insights into diabetic renal injury from the db/db mouse model of type 2 diabetic nephropathy . Am J Physiol Renal Physiol 300 : F301 - 10 .
7. Menzel S , Moeller MJ ( 2011 ) Role of the podocyte in proteinuria . Pediatr Nephrol 26 : 1775 - 80 .
8. Muthukumar T , Dadhania D , Ding R , Snopkowski C , Naqvi R , et al. ( 2005 ) Messenger RNA for FOXP3 in the urine of renal-allograft recipients . N Engl J Med 353 : 2342 - 51 .
9. Wang G , Lai FM , Lai KB , Chow KM , Kwan CH , et al. ( 2008 ) Urinary mRNA expression of ACE and ACE2 in human type 2 diabetic nephropathy . Diabetologia 51 : 1062 - 7 .
10. Zheng M , Lv LL , Ni J , Ni HF , Li Q , et al. ( 2011 ) Urinary podocyte-associated mRNA profile in various stages of diabetic nephropathy . PLoS One 6 : e20431 .
11. Zheng M , Lv LL , Cao YH , Zhang JD , Wu M , et al. ( 2011 ) Urinary mRNA markers of epithelial-mesenchymal transition correlate with progression of diabetic nephropathy . Clin Endocrinol (Oxf). [Epub ahead of print].
12. Roepman P ( 2010 ) The future of diagnostic gene-expression microarrays: bridging the gap between bench and bedside . Bioanalysis 2 : 249 - 62 .
13. Svensson K , Granberg M , Karlsson L , Neubauerova V , Forsman M , et al. ( 2009 ) A real-time PCR array for hierarchical identification of Francisella isolates . PLoS One 4 : e8360 .
14. Munshi R , Johnson A , Siew ED , Ikizler TA , Ware LB , et al. ( 2011 ) MCP-1 gene activation marks acute kidney injury . J Am Soc Nephrol 22 : 165 - 75 .
15. Rodder S , Scherer A , Korner M , Marti HP ( 2011 ) A subset of metzincins and related genes constitutes a marker of human solid organ fibrosis . Virchows Arch 458 : 487 - 96 .
16. Wu CC , Chen JS , Huang CF , Chen CC , Lu KC , et al. ( 2011 ) Approaching biomarkers of membranous nephropathy from a murine model to human disease . J Biomed Biotechnol 581928.
17. Petermann AT , Krofft R , Blonski M , Hiromura K , Vaughn M , et al. ( 2003 ) Podocytes that detach in experimental membranous nephropathy are viable . Kidney Int 64 : 1222 - 31 .
18. Vogelmann SU , Nelson WJ , Myers BD , Lemley KV ( 2003 ) Urinary excretion of viable podocytes in health and renal disease . Am J Physiol Renal Physiol 285 : F40 - 8 .
19. Tam FW , Riser BL , Meeran K , Rambow J , Pusey CD , et al. ( 2009 ) Urinary monocyte chemoattractant protein-1 (MCP-1) and connective tissue growth factor (CCN2) as prognostic markers for progression of diabetic nephropathy . Cytokine 47 : 37 - 42 .
20. Huang Y , Border WA , Yu L , Zhang J , Lawrence DA , et al. ( 2008 ) A PAI-1 mutant, PAI-1R, slows progression of diabetic nephropathy . J Am Soc Nephrol 19 : 329 - 38 .
21. Chawla T , Sharma D , Singh A ( 2010 ) Role of the renin angiotensin system in diabetic nephropathy . World J Diabetes 1 : 141 - 5 .
22. Liu Y ( 2010 ) New insights into epithelial-mesenchymal transition in kidney fibrosis . J Am Soc Nephrol 21 : 212 - 22 .
23. National Kidney Foundation ( 2007 ) KDOQI Clinical Practice Guidelines and Clinical Practice Recommendations for Diabetes and Chronic Kidney Disease . Am J Kidney Dis 49 : S12 - 154 .
24. Ma YC , Zuo L , Chen JH , Luo Q , Yu XQ , et al. ( 2006 ) Modified glomerular filtration rate estimating equation for Chinese patients with chronic kidney disease . J Am Soc Nephrol 17 : 2937 - 44 .