2.1. Tissue samples and RNA isolation
Tumour and normal kidney tissues were obtained immediately after nephrectomy. A homogeneous appearing part of the tumour was divided into two parts, one of them was immediately snap frozen in liquid nitrogene and stored at -80°C and the another part was processed for histology as reference for the frozen material. The rest of tumour and kidney was processed for routine histology. The histological diagnosis was established according to the Heidelberg Classification of Renal Cell Tumours 11
. The diagnosis of samples used for this study was confirmed by SNP- based array analysis 2
. The collection and use of tissue samples, with the patients' previous informed consent, was approved by the Ethics Committee of the University of Heidelberg.
Total RNA from normal and tumour tissues was extracted by TRIzol reagent (Invitrogen GmbH, Karlsruhe, Germany) according to the manufacturer's recommendations. The concentration of RNA was measured with spectrophotometry. The absence of DNA contamination in RNA samples was confirmed by PCR using intron-specific primers for of GAPDH.
2.2. Gene expression profiling and data analysis: pooled samples
We pooled the RNA isolated from five normal kidneys, three Wilms' tumours, three papillary RCC without progression (pT1,G1) and two ones with metastatic growth (pT3,G3), five conventional RCCs without progression (pT1,G1) and four with deadly progression (pT3,G3), and also four chromophobe RCCs and three ROs. Total RNA was purified with Qiagen RNeasy Mini Kit (QIAGEN GmbH, Hilden, Germany). The cRNA synthesis, hybridization onto Affymetrix HG-U133 A and B array set (Affymetrix Inc., Santa Clara, USA) and first data analysis was performed at the German Resource Center for Genome Research (http://www.rzpd.de
). Normalized and log2-transformed intensity data (44,928 probe sets) as well as corresponding present/absent and increased/decreased calls produced by the GCOS software (Affymetrix) were further processed using Microsoft Excel. The value in the pooled normal adult kidney samples was used as a baseline reference and relative expression values in pooled tumour samples were calculated. The genes that showed an adjusted signal log2 ratio above 1 or less than 1 in tumour samples were considered to be over- and under-expressed, respectively.
To compare the expression patterns among renal tumour subtypes, the expression profiles were analyzed by unsupervised hierarchical average linkage clustering algorithm supplied with the microarray data analysis tool TMEV (http://www.tm4.org/mev.html
). The process of similarity measurements was based on the Pearson correlation. To identify genes (approx. 600 probe sets) that correlated significantly with chRCC and RO, a three fold-change threshold versus other tumour subtypes was used in combination with the Affymetrix difference call of increase, marginal increase, decrease and marginal decrease. Finally, 100 significant probe sets were manually selected and the information about the transcripts that they detect was annotated using information obtained from the Affymetrix NETAFFX database (www.affymetrix.com
2.3. Gene expression profiling and data analysis: individual samples
In this setting we used high quality RNAs from 13 conventional RCCs without and 13 with deadly progression, 10 papillary RCCs without and 7 papillary RCCs with progression, four chromophobe RCCs, four ROs, two collecting duct carcinomas, four Wilms' tumours, one clear cell sarcoma of the kidney, one rhabdoid tumour of the kidney, one mucinous tubular and spindle cell carcinoma,
as well as four adult and two fetal normal kidneys. The global gene expression profiling by using the Affymetrix HG-U133 Plus2.0 GeneChip oligonucleotide microarray was carried out at the Genomics Core Facility, EMBL, Heidelberg, Germany. The robust multi-array average algorithm of R and RMA implementation in Bioconductor package (http://www.bioconductor.org
) was used to perform preprocessing of the .CEL files, including background adjustment, quartile normalization, and summarization. Data were further processed using Microsoft Excel. Expression measurements were transformed by computing the base-two logarithm before further analysis. Relative expression profiles were generated from the individual tumour expression profiles and the mean expression values of the four individual normal adult kidney expression profiles (53,677 probe sets).
Differential expression between chromophobe RCCs and ROs versus all other types of RCTs was assessed by ANOVA application supplied by the TMEV tool. Probe sets with P value of < 10-4 were considered significant. To compare the expression patterns among renal tumour subtypes and organize the differentially expressed transcripts the average linkage unsupervised hierarchical cluster analysis (HLC) was performed. The clustering algorithm groups both genes and samples by similarity in expression pattern that was based on the Pearson correlation. The significant analysis of microarrays (SAM) based on two-class unpaired analysis and available number of permutations was used to derive a list of probe sets identified as significantly changed (fold change ≥ 2) between chRCC and RO.
2.4. Functional annotation
Since the interpretation of the microarray results requires the presentation of the data in the context of their functional processes and cellular localization, the set of patterns identified as differentially expressed in chRCC or RO were subgrouped into respective functional categories and according to their cellular localization based on the latest currently available information stored in public domain Database for Annotation, Visualization and Integrated Discovery (DAVID, http://david.abcc.ncifcrf.gov
) and reported GOTERM-BP (Biological process), GOTERM -CC (cellular component) and GOTERM-MF (molecular function) that had corrected P values of <0.01. Although some transcripts have poorly understood functions, functional assignments were possible for most genes.
2.5. Quantative RT-PCR
First strand cDNA synthesis was performed with 2µg of total RNA using SuperScript II reverse transcriptase (Invitrogen) and oligo-p(dT)23
primer according to the manufacturer's instructions. Real-time PCR was performed on a panel of primary RCTs and normal kidneys for assessing the relative expression of genes of interest and validation of microarray expression data. Primers were designed to span at least on intron of genomic sequence using Primer3 software http://frodo.wi.mit.edu/primer3/primer3_code.html
and synthesized by MWG-Biotech AG (Ebersberg, Germany). The sequences of the gene-specific primers used in this study are available upon request. Reaction was performed in 15 µl using 6 μl of cDNA diluted 1:16 and Platinum SYBRGreen qPCR SuperMix-UDG (Invitrogen). After activation at 95 °C for 15 min, PCR product was amplified for 45 cycles (94°C denaturation for 30 s, 60o
C primer annealing for 30 s, and 72°C extension for 45 s) followed by an additional 5 min step at 72 °C. Detection of product amplification, melting curve analysis and semi-quantification was performed by DNA Engine Opticon system (MJ Research, Inc., Watertown, USA). Relative quantity was calculated by dividing the gene specific expression with the expression of β-actin and GAPDH and then brought in correspondence with 100%-scale. All reactions were carried out in duplicates and the results were averaged using statistical analysis tools of MS Excel.
RT-PCR was performed in automated thermal cycler (DNA Engine, Biozym Diagnostik GmbH, Hess.Oldendorf, Germany) in 15 µl reaction volume using 6 µl of cDNA diluted 1:16 and 0.75U Taq polymerase (Invitrogen). In preliminary experiments, cycle titration reactions determined numbers of cycles required for optimal detection was performed. The reaction was done over 30 and 35 cycles; with denaturation at 94°C for 30 sec, annealing at 64°C for 30 sec, elongation at 72°C for 45 sec and an additional 5 min elongation at 72°C. GAPDH was used as the control for equal loading and water as the no template. The expected band for each product was identified by a comigration of a DNA marker ladder electrophoresed in an adjacent lane.
2.7. Western blot analysis
The proteins were solubilized in the lysis buffer (62,5 mM Tris, 2% SDS, 10% sucrose, pH 6,8) containing protease inhibitors (Roche Diagnostics GmbH, Mannheim, Germany), and protein concentrations were measured by BCA Protein Assay kit (Pierce, Rockford, USA). 60 µg of proteins were size-fractioned by 12% SDS-PAGE, transferred electrophoretically to a nitrocellulose membrane (Protran, Whatman GmbH, Dassel, Germany) and blotted with rabbit polyclonal antibody raised against C-terminus of AQP6 (AB3073, Millipore GmbH, Schwalbach, Germany). An enhanced chemiluminescence (Western Lightning Plus-ECL, PerkinElmer LAS GmbH, Rodgau, Germany) system was used for the signal detection. The blot was reprobed with anti-β-actin antibody (AC-15, SIGMA-Aldrich GmbH, Steinhem, Germany) to control for loading.
2.8. Tissue microarray and immunohistochemistry
Original paraffin blocks of fetal and adult kidneys, chromophobe RCCs and ROs as well as tissue microarrays containing fetal and adult kidneys and distinct types of RCTs were used for this study. TMAs with 0,6 mm core biopsies was constructed after marking the areas of interest on H&E stained slides by one of the authors (GK) by a Manual Tissue Arrayer (MTA1, Beecher Instruments, Inc., Sun Prairie, USA).
The 4 µm sections were deparaffinized in xylene, and rehydrated in gradient ethanol. For antigen retrieval, heat pretreatment was performed using the Biocare pressure cooker (10 mM citrate buffer, pH 6.0). The sections were subsequently blocked with 10% normal goat serum (X0907, Dako Cytomation, USA) for 1 h, following by the incubation with primary antibody in a moist chamber at 4oC overnight. After endogenous peroxidase was quenched by 0.3% H2O2 for 15 min, following by incubation with RPH-conjugated secondary antibody (EnVision+ system, Dako North America, Inc., Carpinteria, USA) for 1 h. Visualization of the antigen antibody immunoreaction was completed using AEC (Dako North America, Inc.) for 10 min. The sections were counterstained with hematoxylin, mounted with aqueous permanent mounting medium Ultramount (Dako North America, Inc.) and cover-slipped with Pertex (Medite GmbH, Burgdorf, Germany). The slides were evaluated twice by one of the authors (GK). The following antibodies and dilutions were used: CLDN8 (1:100, GeneTex Inc., Irvine, USA), TMC5 (1:500, Abcam, Heidelberg, Germany), IHPK3 (1:200, Sigma-Aldrich GmbH, Steinhem, Germany), S100A1 (1:200, Acris Antibodies GmbH, Herford, Germany) and AQP6 (1:50, Millipore GmbH, Schwalbach, Germany).