Skip to main content
NCBI home page
The Journal of Molecular Diagnostics : JMD logoLink to The Journal of Molecular Diagnostics : JMD
. 2011 Mar;13(2):143–151. doi: 10.1016/j.jmoldx.2010.10.005

Molecular Markers of Injury in Kidney Biopsy Specimens of Patients with Lupus Nephritis

Heather N Reich ⁎,⁎, Carol Landolt-Marticorena ⁎, Paul C Boutros †, Rohan John ⁎, Joan Wither ⁎, Paul R Fortin ⁎, Stuart Yang ⁎, James W Scholey ⁎, Andrew M Herzenberg ⁎
PMCID: PMC3128621  PMID: 21354048

Abstract

Prediction of prognosis in patients who have lupus nephritis is inadequate, limiting individualization of potentially toxic therapy. Advances in tissue molecular techniques offer new approaches to study mechanisms underlying kidney injury, and add to prognostic information gleaned from biopsy specimens. Analysis of mRNA expression in formalin-fixed, paraffin-embedded renal biopsy specimens is limited by both quantity and quality of RNA, requiring RNA pre-amplification, which can introduce bias. Accordingly, we developed a new technique for RNA extraction from human kidney formalin fixed paraffin embedded biopsy specimens, and used Taqman low-density arrays Applied Biosystems, Carlsbad, CA to simultaneously measure 48 mRNAs in duplicate, in a single biopsy. We extracted mRNA from more than 150 blocks to determine the quantity and vintage of biopsy tissue suitable for analysis using this protocol. We then used Taqman low-density arrays to identify suitable housekeeping genes in lupus nephritis. Finally, we measured expression of 48 mRNA transcripts in archived lupus biopsy specimens (n = 54). We identified that the mRNA levels of three transcripts (MMP7, EGF, COL1A1) relate to pathological indices of kidney injury and kidney function at the time of biopsy; these were associated with parallel changes in expression of these proteins. This new method for measurement of kidney biopsy mRNA expression has enabled us to identify tissue biomarkers of kidney damage and function, and potentially can increase the information yielded from diagnostic kidney biopsy specimens to improve tailoring of therapy.

Renal involvement is common in patients who have systemic lupus erythematosus and clinically evident disease occurs in approximately half of these patients.1,2 Treatment of lupus nephritis necessitates the use of potentially toxic immunosuppressive therapy to prevent progressive tubulo-interstitial scarring and permanent loss of kidney function. The benefits of these medications must be weighed carefully against the potential risks, which include fatal infection, infertility, and late malignancy.3 Clinicians' ability to predict renal prognosis is limited; therefore, the ability to individualize treatment protocols is also inadequate.

A frequently encountered clinical challenge involves the decision whether to press forward with immunotherapy to prevent chronic kidney disease (and accept the attendant risks) or to step down immunotherapy and focus on only conservative treatments (eg, blood pressure management). It is long recognized that serum creatinine is not always a reliable indicator of renal function in patients who have lupus; a substantial decrease in glomerular filtration rate may be required before an increase in serum creatinine levels is observed in this patient population.4 Although renal biopsy may help guide these decisions, it is an invasive procedure, and sampling error can lead to inaccurate estimates of chronic injury. New markers of progressive kidney injury are therefore indicated to allow improved individual tailoring of therapy. The study of tissue molecular markers of fibrosis holds potential for improving clinicians' ability to estimate the extent of chronic injury and predict prognosis.5,6

The study of broad-based gene expression in kidney biopsy specimens for the purposes of investigating pathophysiology of disease or identification of prognostic indicators is not yet widely performed. Although some centers have recently banked tissue in RNA preservatives (with therefore limited clinical follow-up), the study of gene expression in archived formalin-fixed, paraffin-embedded (FFPE) kidney biopsy specimens has been limited by the quantity and quality of tissue and RNA derived from these samples.7–10 Using standard PCR technologies, the expression of only a limited number of gene targets is possible. Pre-amplification has the potential to introduce bias due to uneven RNA degradation. Accordingly, we developed a procedure using modified column-based techniques to extract RNA from routinely archived FFPE 18-gauge renal biopsy specimens that enabled simultaneous analysis of the expression of multiple mRNA transcripts by Taqman real-time RT-PCR low-density arrays (Figure 1). Using this new technique, we determined that the tissue mRNA expression relates to biopsy pathological injury scores and kidney function at the time of biopsy, and transcriptomic changes may also be associated with corresponding translational changes in protein abundance. Our findings provide proof of principle that this approach is a feasible and clinically meaningful method to identify and validate biomarkers of progressive kidney disease.

Figure 1.

Figure 1

Open in a new tab

Experimental workflow for mRNA expression analysis of FFPE 18-gauge renal biopsy specimens. Histological sections of archived FFPE biopsy specimens are cut and placed in a xylene-filled microtube for deparaffinization. After ethanol wash and protease digestion, RNA is isolated. RNA quantity and quality are measured and RNA is reverse transcribed to cDNA. The product is used either in a Taqman low-density array (TLDA) or standard Taqman real-time PCR.

Materials and Methods

Tissue Samples

Biopsy specimens were obtained using an 18-gauge needle and archived after pathological diagnosis. The FFPE blocks were graded: grade 0, no identifiable remaining tissue; grade 1, small tissue fragment approximately <5 mm in length; grade 2, moderate fragment 5 to 10 mm in length; and grade 3, large fragment >10 mm or multiple fragments. For grade 3 biopsy specimens, four 4-μm sections were collected in a 1.5-ml ribonuclease-free tube containing 100% xylene. For lower-grade biopsy specimens, up to six sections were collected.

Deparaffinization

RNA extraction from lupus FFPE biopsy specimens was performed using the RecoverAll Total Nucleic Acid Isolation commercial kit (Applied Biosystems, Carlsbad, CA) with several modifications described below. Samples were vortex-disrupted, heated for 3 minutes at 50°C, and centrifuged; the sample was washed twice with 100% ethanol. The pellet was air-dried for 15 minutes.

Protease Digestion

One hundred microliters of digestion buffer and 6 μL of protease were added and incubated for 6 hours at 50°C.

Nucleic Acid Isolation

One hundred twenty microliters of isolation additive and 275 μL of 100% ethanol were added. The contents were applied to a filter cartridge (RNAqueous-Micro kit, Applied Biosystems). After centrifugation, 500 μL of wash 1 to 3 were sequentially added, and the column was centrifuged.

Nuclease Digestion and Final Nucleic Acid Purification

Fifty microliters of 95°C nuclease-free water, 6 μL of 10x deoxyribonuclease buffer, and 4 μL deoxyribonuclease were applied to the filter, and the column was incubated for 30 minutes. Five hundred microliters of washes 1 to 3 were sequentially performed. Thirty microliters of 95°C nuclease-free water was applied to the filter and incubated for 1 minute. After centrifugation for 1 minute, the RNA was collected. The RNA was analyzed using a NanoDrop 2000 spectrophotometer (Thermo Fischer Scientific).

RNA Extraction for Standard Real-Time PCR Tissue Vintage Experiments

For diabetes samples, conventional real-time PCR was performed. Tissue digestion, RNA extraction, deoxyribonuclease treatment, and first-strand cDNA synthesis were performed using the Paradise Whole Transcript Reverse Transcription kit (Applied Biosystems), using proteinase K buffer and incubation for 16 to 20 hours for digestion. Standardized quantities of tissue were used for RNA extraction. For standard real-time PCR, all measurements were performed in triplicate, and quantified in reference to a 6-dilution standard curve using cDNA obtained from cultured human renal tubular epithelial cells.

Taqman Low-Density Array Real-Time PCR

To ensure that the samples would be compatible for Taqman low-density array (TLDA) analysis, RNA extraction from lupus samples was performed using Applied Biosystems products. We validated similar RNA extraction yields using both kits (data not shown), based on cycle threshold values. Reverse transcription was performed using 100 ng of RNA in 20 μL of solution, according to the manufacturer's directions (High Capacity cDNA kit, Applied Biosystems).

TLDA Housekeeping Gene Card

The cDNA sample (derived from 100 ng of RNA), ribonuclease-free water, and PCR master-mix were loaded into a TLDA-card fill port. The samples were distributed on the plate by centrifugation (Sorvall, Thermo Fisher Scientific, Waltham, MA, USA). We used one preformatted “endogenous control” card with 16 housekeeping genes in triplicate. Real-time PCR was performed on the 7900HT Real-Time PCR system (Applied Biosystems).

TLDA Analysis of mRNA Expression

We designed custom-formatted cards with 48 genes measured in duplicate, including two housekeeping genes. All primer/probe sets were inventoried. Gene expression assays were supplied by Applied Biosystems. The cDNA sample (derived from 100 ng of RNA), ribonuclease-free water, and PCR master-mix were loaded into a TLDA-card fill port.

TLDA Data Analysis

To calculate relative housekeeping gene stability using the preformatted endogenous control TLDA card, the Excel VBA macro “geNorm” was used as described. (http://medgen.ugent.be/∼jvdesomp/genorm).11 This method ranks stability of candidate reference endogenous control genes. To analyze gene expression using the 48-gene TLDA arrays on the remaining 40 samples, the cycle threshold values were normalized as previously described12 and Spearman's correlation was performed.

For the experimental work, TLDA mRNA expression levels were normalized as described previously12 using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and peptidylprolyl isomerase (PPIA, also known as cyclophillin A) as normalization genes (GAPDH standard on plate). Normalized data were associated to clinical and pathological parameters using one-way analysis of variance (Class Analysis), two-tailed t-tests with Welch's adjustment for heteroscedasticity (active versus inactive), and Spearman's correlation using the Best and Robert's method13 for assessing significance (all other clinical covariates). P values from each expression-clinical analysis were separately subjected to a false-discovery rate adjustment for multiple testing. The matrix of −log10(P) across all genes and clinical covariates was subjected to unsupervised machine-learning using the DIANA (DIvisive ANAlysis) clustering algorithm with Pearson's correlation as the distance metric and average linkage, as previously described.14

To determine the predictive value of the expression of all three selected mRNA transcripts on biopsy chronicity, we performed a leave-one-out cross validation with nested feature selection using Spearman's correlation as a metric. The three most correlated features (direction-independent) were fit into a linear model against the chronicity index. This model was then used to predict the held-out patient, and the procedure was repeated for all patients. All analyses were done in the R statistical environment (v2.11.1, http://www.r-project.org).

Immunohistochemistry

FFPE sections from the same renal biopsy were cut at 4 mm followed by heat-induced antigen retrieval. Sections were incubated with monoclonal antibodies directed against the protein of interest (1:400). To eliminate any potential nonspecific biotin activity, slides were stained with a secondary anti-mouse antibody using the EnVision system (Dako Glostrup, Denmark). Endogenous peroxidase activity was prevented by pretreating with 3% hydrogen peroxide. Negative controls with irrelevant primary antibody and no primary antibody were performed. The antibodies tested included: matrix metallopeptidase 7 (MMP7, R&D Systems, Minneapolis, MN), Collagen 1a1 (COL1A1, Sigma, Oakville, ON), and epidermal growth factor (EGF). Semiquantitative scoring of staining intensity was performed on cortical biopsy sections by a pathologist (A.H.) blinded to the antibody and formal pathological score.

Results

Vintage of Biopsy Suitable for Real-Time PCR

We extracted RNA from 120 FFPE biopsy specimens from patients who have diabetic nephropathy and assessed the expression of two housekeeping genes,7 GAPDH and 18s (ribosomal rRNA), to determine vintage of biopsy suitable for PCR studies. The GAPDH and 18s9 mRNA expression were inconsistent in samples from biopsy specimens that have been archived for more than 9 years (data not shown). As shown in Figure 2A, the expression of GAPDH was relatively constant for up to 8 years of archival time, and there was modest variability in expression levels. In contrast, there was a significant decrease in 18s expression in biopsy specimens that were more than 2 years old, and there was high variability in 18s expression in more recent biopsy specimens as shown by time-dependent increase in the coefficient of variation of the mean cycle threshold (Figure 2B).

Figure 2.

Figure 2

Open in a new tab

Effect of archival age on housekeeping gene expression in FFPE diabetes biopsy specimens. Expression of GAPDH and 18s housekeeping genes in 120 archived FFPE biopsy specimens were measured in triplicate using conventional real-time PCR. Mean amplification cycle thresholds (Ct) and SD are shown according to age of biopsy (A; archive time). For 8 years of archival time, the expression of GAPDH was relatively constant, and there was modest variability in expression levels. In contrast, there was a significant time-dependent decrease in 18s expression, and there was high variability in 18s expression in more recent biopsy specimens. B shows the time-dependent variability in 18s expression.

Determination of Tissue Requirement and RNA Yield

Given these results and that our primary experimental interest was lupus nephritis, we used FFPE kidney tissue archived for 9 years or less that had been obtained from patients who were diagnosed with lupus nephritis after the year 2000 (n = 60). Biopsy specimens were graded (see Materials and Methods) according to available tissue, and then sectioned. Visual inspection of the block and biopsy grade was sufficient to determine which biopsy specimens require an additional two sections for digestion (Figure 3). Individual biopsy specimens yielded 16.7 ± 10.7 ng/μL of RNA in a 30-μL elution volume. Use of additional (more than six 4-μm sections) or more than 6-μm-thick sections in a single column extraction did not increase the RNA yield. When RNA purity was assessed using a 260/280 ratio, the larger grade 2 (n = 25) or 3 (n = 25) biopsy specimens yielded higher-quality RNA than grade 1 samples (n = 10, P = 0.05 for analysis of variance comparison across groups). The RNA purity according to 260/230 ratio [indicator of possible carry-over of organic compounds (260/280 and 260/230 ratios)] did not differ according to biopsy size or grade.

Figure 3.

Figure 3

Open in a new tab

Effect of qualitative biopsy grade on RNA quantity and purity. A: RNA concentration. B: RNA nanodrop spectrophotometry 260/280 ratios. C: RNA nanodrop spectrophotometry 260/230 ratios. Values are based on measurements in 60 biopsy samples (grade 1, n = 10; grade 2, n = 25; grade 3, n = 25).

Effect of Biopsy Vintage on RNA Yield and Purity

We divided the lupus FFPE blocks into three groups: 2000 to 2002 (n = 8), 2003 to 2005 (n = 29), and 2006 to 2008 (n = 21) to study the effect of vintage on the 260/280 and 260/230 ratios within this period. RNA concentration was higher in more recent biopsy specimens (P = 0.05) (Figure 4A), and there was a trend toward improved RNA purity according to 260/280 and 260/230 ratios (P = 0.08, Figure 4B; P < 0.05, Figure 4C).

Figure 4.

Figure 4

Open in a new tab

Storage-time effect on RNA concentration and purity in archived lupus FFPE renal biopsy specimens. A: RNA concentration was significantly higher in more recent biopsy specimens (P < 0.05). B: The 260/280 ratio did not differ according to biopsy vintage. C: The 260/230 ratio was superior in the most recent biopsy specimens (*P < 0.05) 2000 to 2002 (n = 8), 2003 to 2005 (n = 29), and 2006 to 2008 (n = 21).

TLDA Analysis

To test the feasibility of TLDA analysis for study of FFPE kidney biopsy specimens, and to select a housekeeping gene that would not vary according to lupus nephritis activity or injury, we analyzed mRNA samples from eight lupus kidney biopsy specimens using a preformatted housekeeping gene array configured to measure 16 housekeeping genes in triplicate (Figure 5 and Supplementary Table S1 at http://jmd.amjpathol.org). These eight biopsy specimens included a wide range of activity and chronicity scores. Expression stability analysis was performed using the geNorm analysis tool.11 As in Figure 5B, cyclophilin A (PPIA) was the most stable housekeeping gene. Surprisingly, the stability of GAPDH gene expression appeared inferior with this assay so we studied the correlation in expression between GAPDH and PPIA in a larger independent set of 40 FFPE lupus biopsy specimens obtained between 2002 and 2008, using a custom-designed TLDA platform that included these two genes (in addition to 46 other transcripts) in duplicate.12 As shown in Figure 5C, the expression of GAPDH and PPIA is highly correlated (r = 0.93, P < 0.05), despite variance in expression, indicating that both of these genes are suitable endogenous controls for lupus biopsy specimens. We performed the remaining analyses using both PPIA and GAPDH as normalizing genes, and the results did not differ.

Figure 5.

Figure 5

Open in a new tab

TLDA housekeeping gene expression lupus biopsy specimens. A: Cycle threshold (Ct) of 16 housekeeping genes measured using the preformatted TLDA housekeeping gene card platform in eight samples in triplicate. A lower cycle threshold indicates higher mRNA abundance. B: Housekeeping gene expression stability using geNorm analysis; reduced gene expression variability compared to other housekeeping genes is reflected by lower score. Cyclophilin A (PPIA) shows the lowest relative variability (most stability). C: Expression of GAPDH compared with cyclophilin A in an additional 40 lupus samples (PPIA), measured using TLDA-based real-time PCR. There is a high correlation between GAPDH and cyclophilin A expression (r = 0.93, P < 0.05).

Relating mRNA Expression to Pathological and Clinical Variables

A set of 54 biopsy specimens from patients who have lupus nephritis were selected for study. The pathological features of the biopsy specimens are provided in Table 1; biopsy specimens exhibited a range of disease activity and chronicity, according to the International Society of Nephrology/Renal Pathology Society classification and scoring system.15 The clinical characteristics of the patients undergoing renal biopsy are also provided in Table 1. The mRNA transcripts measured by TLDA are listed in Supplementary Table S2 (see http://jmd.amjpathol.org); these were selected specifically from the scientific literature to include molecules implicated in the processes of inflammation and fibrosis.

Table 1.

Clinical and Biopsy Characteristics of Patients Who Underwent Diagnostic Kidney Biopsy

Characteristic Class I Class II Class III Class IV Class V (only)
Number 2 6 20 21 5
Median proteinuria at biopsy (g/day) (min, max) 0.11 0.05 (0.01, 0.17) 1.8 (0.04, 7.5) 3.19 (0.54, 3.19) 3.91 (2.6, 8.7)
Mean eGFR at biopsy (ml/min/1.73m2) (min, max) 74 (59, 89) 59 (45, 91) 65 (30, 96) 66 (9126) 88 (31, 120)
Median chronicity index (min, max) 0 1.5 (1, 3) 4 (06) 4 (012) 1 (0, 9)
Open in a new tab

To relate mRNA expression to kidney injury, the normalized mRNA expression levels were related to clinical characteristics of patients at the time of biopsy, and to biopsy chronicity indices. Significant relationships were found to exist between mRNA expression and clinical and pathological variables including kidney function at the time of biopsy, chronicity, and activity scores (Figure 6). By using unbiased statistical approaches (see Materials and Methods), the mRNA transcripts with expression levels most closely related to the disease chronicity index were MMP7, EGF, and COL1A1. Both MMP7 and COL1A1 expression showed a direct relationship to the chronicity index, whereas EGF showed an inverse relationship to chronicity index (Figure 6). We then determined the predictive value of the expression of all three selected mRNA transcripts considered together on biopsy chronicity score. We determined that the chronicity score could be accurately predicted based solely on the expression of these three mRNA transcripts (Rho 0.79, P = 1.78 × 10−12). The expression of MMP7 and EGF were also closely related to the renal function (estimated glomerular filtration rate) at the time of biopsy (Rho −0.39 and 0.34 respectively, unadjusted P value <0.01; see Supplementary Figure S1 at http://jmd.amjpathol.org).

Figure 6.

Figure 6

Open in a new tab

Relationships between mRNA expression and biopsy chronicity scores. A: COL1A1 mRNA. B: MMP7 mRNA. C: EGF mRNA. D: Predicted versus actual chronicity. Analyses are based on mRNA expression measured in duplicate from 54 biopsy specimens.

To determine whether changes in mRNA expression were also associated with changes in protein transcript levels, immunohistochemistry was also performed in the same renal biopsy specimens (n = 19 for MMP7, n = 19 for COL1A1, n = 7 for EGF); representative sections are provided in Figure 7. The biopsy MMP7 (Rho 0.72, P = 0.016), COL1A1 (Rho 0.53, P = 0.02), and EGF (Rho −0.80, P = 0.05) mRNA expression correlated with protein expression and mirrored the relationship to the pathological chronicity index. Biopsy EGF mRNA expression correlated with protein levels, and both EGF mRNA and protein levels showed an inverse relationship with chronicity index.

Figure 7.

Figure 7

Open in a new tab

Representative histological sections and immunohistochemical staining. A and B: There is a direct correlation between COL1A1 protein and mRNA expression, and between COL1A1 protein expression and chronicity score. Biopsy specimens with low chronicity index (A, those with lesser degree of tubulo-interstitial fibrosis) showed low levels of COL1A1 protein expression, and biopsy specimens with a high chronicity index (B) showed high levels of COL1A1 expression. C and D: MMP 7 protein expression is not abundant in biopsy specimens of patients who have low MMP7 mRNA expression and low disease chronicity scores (C); however, there is abundant tubular staining for MMP7 protein in biopsy specimens with high mRNA expression and high chronicity indices (D). E and F: EGF protein expression is low in biopsy specimens with low EGF mRNA expression and high chronicity score (E), but is high in biopsy specimens with high mRNA expression and low chronicity score (F) (direct correlation with mRNA expression, inverse correlation with chronicity score).

The relationship between semiquantitative measurement of protein expression and biopsy chronicity score is shown in Figure 8. As observed at the mRNA level, protein expression of MMP7, COL1A1, and EGF was correlated with chronicity score.

Figure 8.

Figure 8

Open in a new tab

Relationships between protein expression and chronicity index. The correlations between semiquantitative measurement of COL1A1 (A), MMP7 (B), and EGF (C) protein expression and biopsy chronicity scores are shown. Analyses are based on immunohistochemistry staining quantified in biopsy specimens (COL1A1, n = 19; MMP7, n = 19; and EGF, n = 7).

Discussion

In this study we have described the development and validation of a new approach for the study of molecular biomarkers of kidney disease in routinely archived human kidney biopsy specimens. We have used this approach to identify genes/proteins related to pathological indices of kidney injury and kidney function in patients who have lupus nephritis. Although other methods for measuring mRNA expression in kidney biopsy specimens have been described, the advantage of our protocol includes the application in routinely archived samples so that longitudinal clinical data may be available to relate mRNA expression to outcome. Furthermore, our method eliminates bias introduced by pre-amplification techniques required for microarrays, while still allowing simultaneous measurement of multiple transcripts from small samples.

Gene expression analysis by means of quantitative real-time RT-PCR is a valuable technique for genomic studies of kidney disease.7,10,16–18 This approach has already been used to identify tissue biomarkers of cancer prognosis19,20; however, the small amount of kidney biopsy tissue and limited availability of biopsy tissue stored in RNA preservatives has precluded the simultaneous measurement of multiple genes in human kidney biopsy samples in the absence of pre-amplification. Accordingly, we developed a new protocol to simultaneously measure expression of multiple mRNA transcripts in archived FFPE biopsy specimens.

We found that RNA can be extracted from four 4-μm sections from FFPE blocks in sufficient quantity with a simple column-based technique to be used for both standard RT-PCR and TLDA analysis. We were also able to reliably extract RNA from FFPE blocks stored since 2000 (approximately 8 years). Although we were able to extract RNA from older vintage biopsy specimens, the yield of cDNA was variable, precluding reliable analysis by real-time PCR or TLDA. Our observation that increasing the number of tissue sections did not influence the RNA yield merits comment. We hypothesize that this relates to the use of a column-based extraction technique, and that once the surface area of the column is saturated or exceeded, there is no added benefit (in terms of RNA yield or quality) obtained by increasing the quantity of tissue used for nucleic acid extraction.

We measured the expression of 16 housekeeping genes in triplicate on a TLDA platform without a pre-amplification procedure with the RNA extracted by our protocol from eight biopsy specimens. Cycophillin A PPIA) was the most stable housekeeping gene in lupus samples, with minimal variability in expression. Surprisingly, GAPDH expression did not appear to be as stable; however, this may have related to the small number of samples8 studied on the housekeeping gene card. When tested in a larger subset of biopsy specimens, we determined that the expression of PPIA and GAPDH is highly (though not perfectly) correlated. We performed our analyses using PPIA and GAPDH as normalizing genes, and the results did not differ.

Having rigorously established the technical approach, we then measured expression of approximately 50 mRNA transcripts in 54 FFPE biopsy specimens of patients who have lupus nephritis. These transcripts included potential mediators of both inflammation and fibrosis in this disease, drawn from the scientific literature. We used a conservative statistical approach to identify the three transcripts most closely correlated with the kidney biopsy chronicity index. We found that MMP7 and COL1A1 mRNA expression was directly related to the chronicity index, and that EGF expression related inversely to this score (P < 0.05) when corrected for multiple testing. By measuring the expression of all three transcripts together, we were able to closely predict the biopsy chronicity score. We further found that the expression of MMP7 and EGF relate to renal function at the time of kidney biopsy. Finally, we confirmed that protein abundance mirrors mRNA expression for these transcripts (Figures 7 and 8).

One of the hallmarks of renal fibrosis is the elaboration of extracellular matrix by myofibroblast-type cells. The stimulus for the fibrotic response includes infiltration of the kidney parenchyma by macrophages which, along with resident kidney cells, secrete a host of pro-inflamamtory and pro-fibrotic mediators, such as chemokines and growth factors.21 Strong experimental evidence suggests that these stimuli promote epithelial-to-mesenchymal transition of tubular epithelial cells to a fibroblast phenotype, contributing to interstitial fibrosis,22 and elaboration of interstitial matrix components including type I collagen.23 Our finding that the expression of COL1A1 mRNA and protein are positively correlated with the lupus disease chronicity score therefore recapitulates findings in animal models of systemic lupus erythematosus–related renal fibrosis,24 and may be reflective of both chronic injury due to systemic lupus erythematosus–related autoimmune disease and secondary activation of the local renin angiotensin system.25

Matrix metalloproteinases (MMPs) are a large family of matrix-degrading enzymes that are key regulators of extracellular matrix turnover (reviewed in Lenz et al).26 Our finding that MMP7 (matrilysin) expression also correlates with chronicity score supports a potential counterbalance for the increase in collagen expression; an increase in expression of MMPs may reflect a physiological response to counter fibrotic injury. These findings are also in accord with the results of a recent discovery-based analysis of mRNA expression in glomeruli of lupus nephritis biopsy specimens, where increased expression of MMP7 was found in both the peripheral blood and laser-microdissected glomeruli of kidney biopsy specimens from patients who have lupus nephritis.27

EGF expression showed an inverse correlation with chronicity score (Figure 6). EGF serves as a protective trophic factor in the kidney,28 and is involved in renal tubular cell growth and survival and recovery from injury.29 The trophic and anti-apoptotic effects of EGF may also counter the activity of transforming growth factor-β,30 another important mediator of kidney injury in lupus nephritis.31 Urinary excretion of EGF, a surrogate marker of kidney EGF mRNA and protein expression, is reduced in patients who have diabetic nephropathy,32 and in patients with progressive IgA nephropathy.33 One possible future application of our approach is to use tissue mRNA and protein expression to identify new urinary biomarkers of kidney injury; urine EGF, collagen, and MMP7 levels may be noninvasive early markers of kidney fibrosis in patients who have lupus nephritis.

One limitation of this protocol is the use of whole biopsy cores, so that most of the RNA is derived from the tubulo-interstitial compartment of the kidney.

However, tubulo-interstitial injury is one of the final common pathways of glomerular disease, and one of the most important predictors of long-term renal outcome.34–37 Indeed, there was a relatively poor correlation between glomerular pathology and renal function in this group of subjects. Patients who had class II lupus nephritis, for example, had renal insufficiency and evidence of tubulo-interstitial fibrosis (Table 1), which is somewhat unexpected. Therefore, the use of this novel protocol has the potential to identify mechanisms underlying progressive kidney injury and, potentially, functional decline.

In conclusion, we describe a rapid, simple, column-based method for RNA extraction from FFPE kidney biopsy specimens archived for up to 8 years, and the measurement of expression of 48 genes in duplicate using a TLDA platform. This methodology allows investigators to use archived kidney biopsy samples to relate gene expression to clinical and pathological parameters, and identify new mechanisms of progressive kidney injury.

Acknowledgments

The authors gratefully acknowledge the helpful input of Dr. Dafna Gladman and Dr. Murray Urowitz. The Physicians Services Inc. Foundation is supported by the physicians of Ontario.

Footnotes

Supported in part by a grant from the Physicians Services Inc. Foundation; the Canadian Institutes of Health Research (CIHR) New Emerging Team (NET) grant (LunNET) (A.H., P.F., and J.S.); a Canadian Diabetes Association (CDA) grant (A.H. and J.S.); a KRESCENT New Investigator Award (CIHR, Kidney Foundation of Canada, and Canadian Society of Nephrology) (H.R.); and a CIHR-AMGEN Canada Research Chair in Nephrology (J.S.).

The author A.M.H. is deceased.

Supplemental material for this article may be found at http://jmd.amjpathol.org or at doi: 10.1016/j.jmoldx.2010.10.005.

Supplementary data

Supplemental Table S1
mmc1.doc (33.5KB, doc)
Supplemental Table S2
mmc2.doc (67KB, doc)
Supplemental Figure S1
mmc3.pdf (137KB, pdf)

References

  • 1.Cervera R., Khamashta M.A., Font J., Sebastiani G.D., Gil A., Lavilla P., Domenech I., Aydintug A.O., Jedryka-Goral A. Systemic lupus erythematosus: clinical and immunologic patterns of disease expression in a cohort of 1,000 patients: The European Working Party on Systemic Lupus Erythematosus. Medicine (Baltimore) 1993;72:113–124. [PubMed] [Google Scholar]
  • 2.Alarcon G.S., McGwin G., Jr, Petri M., Reveille J.D., Ramsey-Goldman R., Kimberly R.P. Baseline characteristics of a multiethnic lupus cohort: PROFILE. Lupus. 2002;11:95–101. doi: 10.1191/0961203302lu155oa. [DOI] [PubMed] [Google Scholar]
  • 3.Beimler J.H., Andrassy K. Cyclophosphamide treatment in systemic necrotizing vasculitis and lupus nephritis: How long? How much? Pediatr Nephrol. 2004;19:949–955. doi: 10.1007/s00467-004-1553-6. [DOI] [PubMed] [Google Scholar]
  • 4.Shemesh O., Golbetz H., Kriss J.P., Myers B.D. Limitations of creatinine as a filtration marker in glomerulopathic patients. Kidney Int. 1985;28:830–838. doi: 10.1038/ki.1985.205. [DOI] [PubMed] [Google Scholar]
  • 5.Ju W., Eichinger F., Bitzer M., Oh J., McWeeney S., Berthier C.C., Shedden K., Cohen C.D., Henger A., Krick S., Kopp J.B., Stoeckert C.J., Jr, Dikman S., Schroppel B., Thomas D.B., Schlondorff D., Kretzler M., Bottinger E.P. Renal gene and protein expression signatures for prediction of kidney disease progression. Am J Pathol. 2009;174:2073–2085. doi: 10.2353/ajpath.2009.080888. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Izmirly P.M., Barisoni L., Buyon J.P., Kim M.Y., Rivera T.L., Schwartzman J.S., Weisstuch J.M., Liu D.T., Bernstein S., Tseng C.E., Belmont H.M., Esmon C.T., Merrill J.T., Askanase A.D., Thomas D.B., Clancy R.M. Expression of endothelial protein C receptor in cortical peritubular capillaries associates with a poor clinical response in lupus nephritis. Rheumatology (Oxford) 2009;48:513–519. doi: 10.1093/rheumatology/kep034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Cohen C.D., Frach K., Schlondorff D., Kretzler M. Quantitative gene expression analysis in renal biopsies: a novel protocol for a high-throughput multicenter application. Kidney Int. 2002;61:133–140. doi: 10.1046/j.1523-1755.2002.00113.x. [DOI] [PubMed] [Google Scholar]
  • 8.Kretzler M., Cohen C.D., Doran P., Henger A., Madden S., Grone E.F., Nelson P.J., Schlondorff D., Grone H.J. Repuncturing the renal biopsy: strategies for molecular diagnosis in nephrology. J Am Soc Nephrol. 2002;13:1961–1972. doi: 10.1097/01.asn.0000020390.29418.70. [DOI] [PubMed] [Google Scholar]
  • 9.Schmid H., Cohen C.D., Henger A., Irrgang S., Schlondorff D., Kretzler M. Validation of endogenous controls for gene expression analysis in microdissected human renal biopsies. Kidney Int. 2003;64:356–360. doi: 10.1046/j.1523-1755.2003.00074.x. [DOI] [PubMed] [Google Scholar]
  • 10.Schmid H., Cohen C.D., Henger A., Schlondorff D., Kretzler M. Gene expression analysis in renal biopsies. Nephrol Dial Transplant. 2004;19:1347–1351. doi: 10.1093/ndt/gfh181. [DOI] [PubMed] [Google Scholar]
  • 11.Vandesompele J., De Preter K., Pattyn F., Poppe B., Van Roy N., De Paepe A., Speleman F. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002;3:1–34. doi: 10.1186/gb-2002-3-7-research0034. research0034.11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Hui A.B.Y., Shi W., Boutros P.C., Miller N., Pintilie M., Fyles T., McCready D., Wong D., Gerster K., Jurisica I., Penn L.Z., Liu F.F. Robust global micro-RNA profiling with formalin-fixed paraffin-embedded breast cancer tissues. Lab Invest. 2009;89:597–606. doi: 10.1038/labinvest.2009.12. [DOI] [PubMed] [Google Scholar]
  • 13.Best D.J., Roberts D.E. Algorithm AS 89: upper tail probabilities of Spearman's Rho. Applied Statistics. 1974;24:377–379. [Google Scholar]
  • 14.Boutros P.C., Yan R., Moffat I.D., Pohjanvirta R., Okey A.B. Transcriptomic responses to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in liver: comparison of rat and mouse. BMC Genomics. 2008;9:419. doi: 10.1186/1471-2164-9-419. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Weening J.J., D'Agati V.D., Schwartz M.M., Seshan S.V., Alpers C.E., Appel G.B., Balow J.E., Bruijn J.A., Cook T., Ferrario F., Fogo A.B., Ginzler E.M., Hebert L., Hill G., Hill P., Jennette J.C., Kong N.C., Lesavre P., Lockshin M., Looi L.M., Makino H., Moura L.A., Nagata M. The classification of glomerulonephritis in systemic lupus erythematosus revisited. J Am Soc Nephrol. 2004;15:241–250. doi: 10.1097/01.asn.0000108969.21691.5d. [DOI] [PubMed] [Google Scholar]
  • 16.Henger A., Kretzler M., Doran P., Bonrouhi M., Schmid H., Kiss E., Cohen C.D., Madden S., Porubsky S., Grone E.F., Schlondorff D., Nelson P.J., Grone H.J. Gene expression fingerprints in human tubulointerstitial inflammation and fibrosis as prognostic markers of disease progression. Kidney Int. 2004;65:904–917. doi: 10.1111/j.1523-1755.2004.00499.x. [DOI] [PubMed] [Google Scholar]
  • 17.Schmid H., Boucherot A., Yasuda Y., Henger A., Brunner B., Eichinger F., Nitsche A., Kiss E., Bleich M., Grone H.J., Nelson P.J., Schlondorff D., Cohen C.D., Kretzler M. Modular activation of nuclear factor-kappaB transcriptional programs in human diabetic nephropathy. Diabetes. 2006;55:2993–3003. doi: 10.2337/db06-0477. [DOI] [PubMed] [Google Scholar]
  • 18.Rudnicki M., Eder S., Perco P., Enrich J., Scheiber K., Koppelstatter C., Schratzberger G., Mayer B., Oberbauer R., Meyer T.W., Mayer G. Gene expression profiles of human proximal tubular epithelial cells in proteinuric nephropathies. Kidney Int. 2007;71:325–335. doi: 10.1038/sj.ki.5002043. [DOI] [PubMed] [Google Scholar]
  • 19.O'Shaughnessy J.A. Molecular signatures predict outcomes of breast cancer. N Engl J Med. 2006;355:615–617. doi: 10.1056/NEJMe068145. [DOI] [PubMed] [Google Scholar]
  • 20.Hoshida Y., Villanueva A., Kobayashi M., Peix J., Chiang D.Y., Camargo A., Gupta S., Moore J., Wrobel M.J., Lerner J., Reich M., Chan J.A., Glickman J.N., Ikeda K., Hashimoto M., Watanabe G., Daidone M.G., Roayaie S., Schwartz M., Thung S., Salvesen H.B., Gabriel S., Mazzaferro V., Bruix J., Friedman S.L., Kumada H., Llovet J.M., Golub T.R. Gene expression in fixed tissues and outcome in hepatocellular carcinoma. N Engl J Med. 2008;359:1995–2004. doi: 10.1056/NEJMoa0804525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Ricardo S.D., van Goor H., Eddy A.A. Macrophage diversity in renal injury and repair. J Clin Invest. 2008;118:3522–3530. doi: 10.1172/JCI36150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Liu Y. New insights into epithelial-mesenchymal transition in kidney fibrosis. J Am Soc Nephrol. 2010;21:212–222. doi: 10.1681/ASN.2008121226. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kalluri R., Neilson E.G. Epithelial-mesenchymal transition and its implications for fibrosis. J Clin Invest. 2003;112:1776–1784. doi: 10.1172/JCI20530. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Anders H.J., Belemezova E., Eis V., Segerer S., Vielhauer V., Perez de Lema G., Kretzler M., Cohen C.D., Frink M., Horuk R., Hudkins K.L., Alpers C.E., Mampaso F., Schlondorff D. Late onset of treatment with a chemokine receptor CCR1 antagonist prevents progression of lupus nephritis in MRL-Fas(lpr) mice. J Am Soc Nephrol. 2004;15:1504–1513. doi: 10.1097/01.asn.0000130082.67775.60. [DOI] [PubMed] [Google Scholar]
  • 25.Mezzano S.A., Ruiz-Ortega M., Egido J. Angiotensin II and renal fibrosis. Hypertension. 2001;38:635–638. doi: 10.1161/hy09t1.094234. [DOI] [PubMed] [Google Scholar]
  • 26.Lenz O., Elliot S.J., Stetler-Stevenson W.G. Matrix metalloproteinases in renal development and disease. J Am Soc Nephrol. 2000;11:574–581. doi: 10.1681/ASN.V113574. [DOI] [PubMed] [Google Scholar]
  • 27.Peterson K.S., Huang J.F., Zhu J., D'Agati V., Liu X., Miller N., Erlander M.G., Jackson M.R., Winchester R.J. Characterization of heterogeneity in the molecular pathogenesis of lupus nephritis from transcriptional profiles of laser-captured glomeruli. J Clin Invest. 2004;113:1722–1733. doi: 10.1172/JCI19139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Fisher D.A., Salido E.C., Barajas L. Epidermal growth factor and the kidney. Annu Rev Physiol. 1989;51:67–80. doi: 10.1146/annurev.ph.51.030189.000435. [DOI] [PubMed] [Google Scholar]
  • 29.Gilbert R.E., Cox A., McNally P.G., Wu L.L., Dziadek M., Cooper M.E., Jerums G. Increased epidermal growth factor in experimental diabetes related kidney growth in rats. Diabetologia. 1997;40:778–785. doi: 10.1007/s001250050749. [DOI] [PubMed] [Google Scholar]
  • 30.Teteris S.A., Menahem S.A., Perry G., Maguire J.A., Dowling J.P., Langham R.G., Thomson N.M., Stein A.N. Dysregulated growth factor gene expression is associated with tubulointerstitial apoptosis and renal dysfunction. Kidney Int. 2007;71:1044–1053. doi: 10.1038/sj.ki.5002176. [DOI] [PubMed] [Google Scholar]
  • 31.Ferraccioli G., Romano G. Renal interstitial cells, proteinuria and progression of lupus nephritis: new frontiers for old factors. Lupus. 2008;17:533–540. doi: 10.1177/0961203307088002. [DOI] [PubMed] [Google Scholar]
  • 32.Mathiesen E.R., Nexo E., Hommel E., Parving H.H. Reduced urinary excretion of epidermal growth factor in incipient and overt diabetic nephropathy. Diabet Med. 1989;6:121–126. doi: 10.1111/j.1464-5491.1989.tb02098.x. [DOI] [PubMed] [Google Scholar]
  • 33.Torres D.D., Rossini M., Manno C., Mattace-Raso F., D'Altri C., Ranieri E., Pontrelli P., Grandaliano G., Gesualdo L., Schena F.P. The ratio of epidermal growth factor to monocyte chemotactic peptide-1 in the urine predicts renal prognosis in IgA nephropathy. Kidney Int. 2008;73:327–333. doi: 10.1038/sj.ki.5002621. [DOI] [PubMed] [Google Scholar]
  • 34.Remuzzi G., Bertani T. Pathophysiology of progressive nephropathies. N Engl J Med. 1998;339:1448–1456. doi: 10.1056/NEJM199811123392007. [DOI] [PubMed] [Google Scholar]
  • 35.Schainuck L.I., Striker G.E., Cutler R.E., Benditt E.P. Structural-functional correlations in renal disease: II. The correlations. Hum Pathol. 1970;1:631–641. doi: 10.1016/s0046-8177(70)80061-2. [DOI] [PubMed] [Google Scholar]
  • 36.Rodriguez-Iturbe B., Johnson R.J., Herrera-Acosta J. Tubulointerstitial damage and progression of renal failure. Kidney Int. 2005;68(Suppl 99):S82–S86. doi: 10.1111/j.1523-1755.2005.09915.x. [DOI] [PubMed] [Google Scholar]
  • 37.Gilbert R.E., Cooper M.E. The tubulointerstitium in progressive diabetic kidney disease: more than an aftermath of glomerular injury? Kidney Int. 1999;56:1627–1637. doi: 10.1046/j.1523-1755.1999.00721.x. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Table S1
mmc1.doc (33.5KB, doc)
Supplemental Table S2
mmc2.doc (67KB, doc)
Supplemental Figure S1
mmc3.pdf (137KB, pdf)

Articles from The Journal of Molecular Diagnostics : JMD are provided here courtesy of American Society for Investigative Pathology

RESOURCES