Animals care/tissue collection/dosing
All animal studies were approved by the Institutional Animal Care and Use Committee of Wyeth Research, Collegeville, PA and were conducted in accordance with Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC) guidelines. Sprague Dawley rats (Taconic, Germantown NY) were OVX at 8 weeks of age and rested for about 1 week before administration of daily subcutaneous doses of 17β-estradiol (20 μg/kg) or vehicle (50% DMSO/50% 1× Dulbecco's PBS). Animals were euthanized after 6 hrs, 3 days and 5 days of treatment. For the 3- and 5-day timepoints, animals were euthanized 6 hours after the last dose. At necropsy, the vaginal vault was excised and trimmed of adherent tissue (e.g. the urethra). Typically, organs were trimmed just below the cervix and above the external orifice of the vagina before flash freezing in liquid nitrogen.
For the studies using betacellulin, 8-week-old female Sprague Dawley rats were OVX and allowed to recover from surgery for 10 days prior to treatment. Rats were dosed intravaginally for 5 days with mouse betacellulin (200 nM; R&D Systems, Minneapolis MN) once daily using a carbomer-934P-based vehicle (pH 7.0) and a 0.5 inch long, 18 gauge ball-tipped gavage needle. Following euthanasia by CO2 asphyxiation and pneumothorax, the vaginal vault was removed, flattened on an index card and fixed in 10% buffered formalin (VWR International, West Chester, PA) for 16–72 hours. Vaginas were equilibrated in 15% sucrose for 3 hours prior to embedding in OCT (Sigma, St Louis MO) and cryosectioning. Ten micron sections were stained with hematoxylin and eosin (both from Sigma) for histological analysis.
Human tissue collection
A clinical study was initiated to collect 3 mm vaginal biopsy specimens from post-menopausal women clinically diagnosed with VA. The diagnosis of VA was confirmed by a vaginal maturation index (VMI) containing ≤5% superficial cells and vaginal pH > 5.0. Patients were treated for 3 months with 17β-estradiol, delivered via a skin patch, at a dose of 0.05 mg/day. All 19 subjects had increased VMI (>5%) and/or reduced pH (< or = 5) following treatment. Vaginal biopsies were collected on day 1 prior to the initiation of treatment with 17β-estradiol and on the last day of treatment (day 85). Samples were immediately frozen on dry ice and stored at -80°C until RNA isolation. This study was conducted at the Ohio State University, Division of Clinical Pharmacology (Columbus, Ohio). Written informed consent was obtained from all subjects prior to any study-related procedure. Western Institutional Review Board (Olympia, WA) approved the protocol and informed consent. The details of the clinical analyses have been reported elsewhere .
Each rat vagina was homogenized in 3 mL of TRIzol reagent (Invitrogen, Carlsbad CA) using a Polytron PT3000 grinder (Brinkmann, Westbury NY) at a setting of 12–15 for 1 min. Subsequent RNA extraction was as described previously . For human samples, RNA extraction was performed according to a modified RNeasy mini kit method. (Qiagen, Valencia, CA) .
ComplementaryDNA was amplified from 50 ng total RNA with the Ovation RNA Amplification System (Nugen, San Carlos CA). Real-time PCR reactions (qRT-PCR) were conducted using the TaqMan® Universal PCR Master Mix (Applied Biosystems). Forward and reverse primers (10 μM) and 20 μM labeled probe was mixed with 2× PCR master mix in a total volume of 15 ul and added to 50 ng cDNA in final reaction volume of 25 μl.
PCR reactions were run on an automated fluorometer (ABI Prism 7000 Sequence Detection System, Applied Biosystems) in a 96-well format. PCR conditions for all reactions were as follows: 1 cycle of 50°C for 2 min., 95°C for 10 min. followed by 40 cycles of 95°C for 15 sec. and 60°C for 1 min. Relative expression was determined by using the CT method (Pontius et al., 2003) using Sequence Detector software, version 1.6.3. Complementary DNAs of each sample were evaluated in triplicate. Results were normalized to a control gene, SFSR10 RNA, which has been shown to have very consistent expression across all patient samples by microarray analysis .
The primer sequences for each transcript are as follows:
BTC: forward primer, 5'-GGTGTGAGAGAGTTGACTTGTTTTACC-3'; reverse primer, 5'-TGCTATCAAACAAATCACCAGAATC-3'; probe, 5'-TGTCCTCTGTCTCCTCT-3'
SFSR10: forward primer, 5'-AAGGGAACACTATACCTGTCATGGA-3'; reverse primer, 5'-GGTAAGCAAAGGACCTGAAAATAATATT-3'; probe, 5'-AAGACTTTGCCTGTTCAT-'
For rat vaginal samples, 5 μg of total RNA were used to generate biotin labeled cRNA using an oligo T7 primer in a reverse transcription reaction followed by in vitro transcription reaction with biotin labeled UTP and CTP. Ten μg of cRNA were fragmented and hybridized to RAE230A and RAE230B arrays (Affymetrix, Santa Clara, CA). For human biopsies, RNA was labeled using Ovation™ Biotin System (Nugen, San Francisco, CA). Labeled cDNA (2.5 μg) was fragmented and hybridized to HGU133 Plus 2.0 arrays (Affymetrix, Santa Clara, CA) . Hybridized arrays were stained according to manufacturer's protocols on a Fluidics Station 450 and scanned on an Affymetrix scanner 3000. All array images were visually inspected for defects and quality. Arrays with excessive background, low signal intensity, or major defects within the array were eliminated from further analysis. Signal values were determined using Gene Chip Operating System 1.0 (GCOS, Affymetrix). For each array, all probe sets were normalized to a mean signal intensity value of 100. The default GCOS statistical values were used for all analysis. Signal values and absolute detection calls were imported into Genesis 2.0 (GeneLogic, Gaithersburg, MD) for analysis. The complete data set is available at the Gene Expression Omnibus as GSE11622 .
Determination of probe sets corresponding to homologous gene pairs
Mapping of RAE230 to HGU133 plus 2.0 probe sets was based on NCBI's Homologene (clusters of orthologous genes). First, all qualifiers were annotated with unigene accession numbers through BioExpress (GeneLogic). These accession numbers then were used to map orthologous sequences through NCBI Homologene . Commonly one probe set from one species maps to multiple probe sets from a second species. To resolve multiple mappings, clusters of orthologous probe sets were generated as follows. For each human probe set, all corresponding rat probe sets were identified and grouped into a "cluster". Human probe sets that map to the same set of rat probes were clustered. For each human cluster, the corresponding locuslink ID(s) were identified. Clusters with multiple Locuslink ID annotations were discarded. Approximately 10% of the human probe sets clustered to multiple Locuslink IDs. Clustering identified 9093 groups, spanning 22460 human probe set IDs and 13219 rat probe set IDs. Each cluster was annotated (assigned GO/KEGG etc categories) based on the human Locuslink ID in that cluster.
Identification of regulated genes
Identification of estrogen-regulated genes was done in two independent ways. In the first method, probe sets were considered estrogen regulated if the following conditions were met: 1) A gene was expressed greater than 50 signal units, 2) the percentage of samples with a Present (P) call as determined by GCOS default settings was greater than or equal to 66% in any sample condition, 3) the fold change between all estrogen treated and vehicle treated samples was at least 1.7, and 4) the p-value based on an Wilcoxon t-test was ≤0.01 (rat) or based on a Wilcoxon paired test was <0.01 (human). Over 5300 rat probe sets and 1668 human probe sets met these conditions. The second method used the 9093 clusters identified above. Arrays were subjected to a Loess transformation and the CyberT-statistics were calculated. False discovery rate estimates (q-values) were estimated using a null distribution of CyberT's from permuted data.
Identification of significantly regulated gene sets
Statistical identification of significantly regulated biological pathways was implemented using a modified approach of Tian et al. . 7798 gene sets were tested in the sigpathway analysis. These were derived from the follow sources: 1) GO annotations (from Affymetrix) , 2) KEGG/GenMAPP annotations (from Affymetrix) , 3) Functional categories (from MSigDB), 4) Genes that are located in close proximity along the chromosome (from MSigDB), 5) Genes containing instances of a regulatory motif of interest (from MSigDB), and 6) Genes that are co-regulated with a given gene in a variety of datasets (from MSigDB). Gene sets from MSigDB were derived from pathway databases (Biocarta, KEGG, and BioCyc) and pathway-specific microarray annotations , as well as >5,000 gene sets from Gene Ontology. Only gene sets in which the number of genes was greater or equal to 10 and less than or equal to 400 were used for further analysis. Each gene set was assigned a score of differential expression between two conditions (essentially, the sum of t-statistics for each gene within the set), and the false discovery rate (Q) associated with this score was calculated via two separate permutation tests (permuting gene labels (Q1) and sample labels (Q2)). A gene set was considered significant if its two false discovery rates are below 0.15 (Q1 <= 0.15 and Q2 <= 0.15) for all 4 comparisons. 45 gene sets passed these criteria.