Brown Norway rats were obtained from Charles River Laboratories (L'Arbresle, France) or Elevage Janvier (Le Genest-Saint-Isle, France). Female rats with their progeny (3 to 5 days old) were purchased from Charles River Laboratories. From their arrival until the time at which they were killed, the rats were kept in the animal house (12:12 h light/dark cycles with free access to food and water) under the rules established by the Board of the Ethical Committee.
Olfactory epithelium dissection
Rats were anesthetized with an injection of 0.3 ml/100 g body weight ketamine hydrochloride (Clorketam 1000 from Vetoquinol). They were then killed by decapitation. Rat skulls were opened through a sagittal section and right and left olfactory epithelia were quickly removed and placed separately in RA1 Buffer from the Nucleospin RNA II kit (Macherey-Nagel, Düren, Germany).
Total RNA was isolated with the Nucleospin RNA kit, according to the manufacturer's (Macherey-Nagel, Düren, Germany) instructions, which included an in-column DNase treatment before RNA elution, to ensure the absence of genomic DNA. Recovered RNA was quantified with a Nanodrop ND-1000 spectrophotometer (NanoDrop Technologies, Cambridge, UK), and RNA integrity was assessed with the RNA 6000 Nano LabChip kit, using the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto CA, USA). Only RNA samples with an RNA Integrity Number (RIN) greater than 8.8 were used for further analysis (RNA profiling analysis and real time reverse transcription PCR analysis). Application of this strict quality threshold resulted in the elimination of the left sample from one male adult rat from Elevage Janvier.
Target preparation and microarray hybridization
RNA samples were labeled with the Agilent Low RNA Input Fluorescent Linear Amplification kit (p/n 5184-3523), according to the manufacturer's instructions. Briefly, 350 ng of total RNA was used as template for reverse transcription to generate cDNA, which was transcribed with T7-polymerase; cyanine-3 (Cy3)-labeled CTP was used for labeling. Cy3 labeling was monitored with a Nanodrop ND-1000 spectrophotometer and was found to be between 1.2 and 1.9 pmol/μl.
Hybridization was performed with the Agilent Gene Expression Hybridization kit (p/n 5188-5242), used according to the manufacturer's instructions. Briefly, 1650 ng of labeled cRNA from each RNA sample was mixed with Hybridization Buffer and Blocking Agent and subjected to fragmentation (by incubation for 30 min at 60°C in the dark). Hybridizations onto 4 × 44K Whole Rat Genome 60-mer oligonucleotide microarrays (G4131F) (Agilent Technologies, Palo Alto CA, USA) were performed in a rotary oven (65°C, 17 h and 10 rpm) in the dark. Slides were disassembled and washed in Gene Expression Wash Buffers I and II, according to the manufacturer's instructions, and dried with a nitrogen-filled air gun before scanning. Fifteen arrays were used for the experiment analyzing the male/female and rat origin comparisons. Thirty-five arrays were used for the "aging" comparison.
Data acquisition and processing
Microarrays were scanned with a dynamic autofocus microarray scanner (Agilent DNA Microarray Scanner), using Agilent parameters. Feature Extraction software version 9.5 was used to extract and analyze the signals. Array results were analyzed with GeneSpring GX software version 7.3 (Agilent Technologies), via the Enhanced Agilent Feature Extraction Import Preprocessor.
Data were normalized in two steps, using the algorithms supplied with the Feature Extraction software. Data were first transformed to convert any negative value to 0.01; normalization was then performed by a per-chip 50th percentile method, which normalized the data for each chip with respect to the median of the chip concerned, allowing comparison between chips. A second normalization step was applied to the results for each gene across all the arrays in the study ("normalize to median"): the median of all the values obtained for a given gene was calculated and used as the normalization standard for that gene, such that, regardless of absolute differences in the expression of the various genes, all were placed on the same scale for comparison.
The accuracy of microarray results was assessed by comparing the overall gene expression levels for each chip by box plot analysis. Each box plot was centered on zero, with comparable dynamic intensities, demonstrating the technical homogeneity of the experiment overall (data not shown).
The microarray data have been uploaded into the Gene Expression Omnibus (GEO) database (SuperSeries no. GSE15954 and samples nos. GSM400094-GSM400143).
Low-intensity and unreliable results were removed using a "filter on flags" feature, with standardized software algorithms classifying spots as "present," "marginal," or "absent". Spots were considered "present" only if the output was uniform, not saturated and significantly above background; spots that satisfied the main requirements but were outliers relative to the typical values for the other genes were considered "marginal". Filters were set to retain for further analysis only those values classified as "present" or "marginal".
The terms "present" or "marginal" defining the nature of the hybridization signals on each microarray should not be confused with the terms expressed transcripts, weakly expressed transcripts and not expressed transcripts defined by comparing the results obtained with the different samples, as explained in Figure .
Content of the 44K Agilent microarrays
There are currently 44,012 probes on each microarrays. By annotation assignments [45
], accession numbers could be assigned to 39,308 of the 39,688 probes for which the manufacturer provided chromosomal location information: rat GenBank accession numbers were assigned for 36,383 of the probes; rat Ensembl transcript identifications (IDs) were assigned for 168 other probes; and non-rat accession numbers for 2,757 probes for which no rat annotations were available. Together, these probes encompass 23,642 unique rat accession numbers and 2,270 unique non-rat accession numbers and represent 16,947 rat Unigene IDs plus 5,941 non-rat Unigene IDs (Unigene build 166). In addition to these probes, there are a number of so-called technical probes engineered by Agilent and used by GeneSpring to ascertain the quality of the data. For additional details, please consult the Agilent website [46
Due to uncertainties regarding the names of a number of genes that are probed by many oligonucleotides on the arrays, the term "gene transcripts", used throughout this paper, designates transcripts and genes collectively identified by these probes, except for OR genes that are annotated as such. Although some gene transcripts were probed by more than one oligonucleotide, each OR gene was probed by a single oligonucleotide.
Selection of differentially expressed genes
We performed t-test analysis with GeneSpring software (Benjamini & Hochberg correction for false discovery rate (p value of 0.01)) to select genes that were differentially expressed between groups.
Hierarchical support trees including bootstrap analysis with replacement after 1000 iterations were constructed with TIGR Mev v 4.2 software [47
]. Numbers at the nodes (range = 1 to 100) indicate the support for the clustering. The clustering pattern was generated by Pearson Correlation with average linkage clustering.
Analysis of the enrichment of expressed genes with Gene Ontology (GO) categories (i.e. GO terms with a significantly larger number of associated genes than expected for a random distribution) was performed with NIH DAVID [48
]. Briefly, the GenBank accession numbers of the genes of interest were uploaded to the DAVID website and analysis was carried out with the Rattus norvegicus
gene repertoire as a reference list. GO categories with significantly larger numbers of expressed genes than expected (p
value corrected < 0.05) were selected.
Real time reverse transcription PCR analysis (RTqPCR)
RTqPCR was performed for a number of genes, with forward (F) and reverse (R) primers designed with Primer3 software [50
] (additional file 8
). Primer specificity was assessed from the monophase dissociation curves. Only pairs presenting similar efficiencies (100 ± 5%) were retained (data not shown). Briefly, the High-Capacity cDNA Archive kit (Applied Biosystems, Foster City, CA, USA) was used for reverse transcription and the Power SYBR Green PCR master kit (Applied Biosystems) was used for quantitative PCR, according to Applied Biosystems gene amplification specifications (40 cycles of 15 s at 95°C and 1 min at 60°C). Gene expression was analyzed with the ABI Prism 7900HT sequence detection system, and results were handled with the associated SDS version 2.3 software (Applied Biosystems).
(hypoxanthine-guanine phosphoribosyltransferase) mRNA levels did not vary significantly between groups or experiments. This gene was therefore used as an internal reference for the comparison of rats of different origins and ages. The relative amounts of gene transcripts were determined by the Ct method [51
]. Each PCR was carried out in triplicate. Results from different samples were compared to a "control sample" corresponding to RNA prepared from one adult rat epithelium.