Indian J Physiol Pharmacol 2010; 54 (3) : 235–254
SMALL SCALE TRANSCRIPT EXPRESSION PROFILE OF HUMAN FIRST TRIMESTER PLACENTAL VILLI ANALYZED BY A CUSTOM-TAILORED cDNA ARRAY M. A. KHAN 1 , M. KAR 1 , S. MITTAL 2 , S. KUMAR 2 , V. L. BHARAGAVA 3 , J . S E N G U P T A 1 A N D D . G H O S H 1* Departments of 1 Physiology and 2 Obstetrics & Gynecology, All India Institute of Medical Sciences, New Delhi – 110 029 and 3
Department of Obstetrics & Gynecology, Sitaram Bhatia Institute of Science & Research, New Delhi – 110 016 ( Received on March 25, 2010 ) Abstract : Human placental trophoblastic mass grows rapidly between 4 and 8 weeks of gestation making it highly vulnerable to external and internal challenges, however, there has been no reported study exploring the developmental molecular characteristics in human first trimester placental villi. In the present study, transcript expressions of human placental villi of normal pregnancies during 6 (n=6), 7 (n=6) and 8 (n=6) weeks of gestation using custom-tailored cDNA-based expression arrays for ~400 annotated human gene products were examined. Unsupervised and supervised analyses of expression data revealed that 386 (95%) genes were overtly involved in the first trimester placental villi, and these genes segregated into three clusters specifically corresponding to 6-, 7- and 8weeks of gestation in principal component analysis. Bayesian prediction analysis based on relative expression levels of genes studied identified that expression patterns in 15 samples out of 18 samples showed concordance with high (0.8-1.0) confidence measures with the chronological age of the placenta, however, two samples collected during 7-weeks of gestation and one sample collected during 8-weeks of gestation were predicted to be 6weeks sample with confidence measures between 0.6 and 0.5. Unsupervised hierarchical clustering analysis segregated the samples into two major branches; while one of them was composed of five 7-weeks samples only, the second major branch had three sub-branches: one of them was exclusively composed of three 8-week samples only, while other two subbranches were mainly composed of 6-weeks samples. K-means clustering analysis identified four optimal clusters of genes depending on the similarity of their relative expression for the set of genes studied across all the samples. Gene ontology (GO) based functional classifications of genes in K-means clusters revealed that the overall putative functions of co-regulated gene clusters were mutually comparable, however, specific genes related to ion homeostasis, metabolism, and VEGF activity specifically clustered in 8-weeks samples. Analysis of relative gene expression during in 6-8 week
* Corresponding Author :
email:
[email protected]
236
Khan et al
Indian J Physiol Pharmacol 2010; 54(3)
placental villi revealed that a large number of gene products were over represented by their either up-regulation (70 genes: ~18%) or down regulation (53 genes; ~14%) between 6 and 8 weeks villi samples and these genes are reportedly involved in biological processes like regulation of cell growth and proliferation, anti-apoptosis, angiogenesis, immune and inflammatory responses, extracellular matrix remodeling and multicellular organismal development involving almost all cellular components and molecular functions like signal transduction activity, transcription factor activity, nucleotide and protein binding, ion (especially calcium and zinc) binding and growth receptor activities. Interestingly, four genes (oxytocin receptor, tenascin C, TNF-R1 and retinol binding protein 1) showed differential regulation in human placental villi during 6-8 weeks of gestation, suggestive of an underlying network of regulation involving these factors in the developing placenta. To our knowledge, this is the first report indicating that these genes are involved in the early stage development of human placenta. K e y w o r d s : early placenta villi
INTRODUCTION In the human, the placental trophoblastic mass grows rapidly between 4 and 8 weeks of gestation accompanied by increased synthesis and secretion of several hormones (1, 2). This time period is highly vulnerable to external and internal trauma that may result in dysregulated early placental development in women associated with high (50-70%) incidence of pregnancy loss during the first trimester, and also with increased risk of developing several types of complications during later stages of pregnancy like preeclampsia, intrauterine growth retardation and abortion (3, 4). We have earlier reported that human placental villi during 6- to 8-week of gestation displayed a complex pattern of homeodynamics involving proliferation, development associated syncytialization and apoptosis (5). Although there a few recent reports wherein human first trimester placental villi have been examined using high throughput platform of transcriptomics and proteomics in reference to several placental pathology (6-8), there has been no reported study exploring the developmental molecular characteristics of
gene expression
transcriptomics
timed human first trimester placental trophoblast cells. In the present study, we examined the transcript expression of human placental villi of normal pregnancies during 6- to 8-weeks of gestation using customtailored cDNA-based expression arrays for ~400 gene products assumed to be important in placental functions during early pregnancy. We are reporting for the first time that a cohort of genes (386) of placental villi displayed differential expression and regulation in specific manner during 6-8 weeks of gestation. Furthermore, we have examined the custom cDNA based expressional information in terms of putative functional significance in developing placental villi. MATERIALS Tissue
AND
METHODS
samples
First trimester human placental samples were obtained from twenty four healthy proven fertile women within age group of 23–30 years undergoing elective surgical termination of pregnancy during 6 to 8 weeks of gestation without any prior medication. The gestational age was estimated from menstrual history, physical
Indian J Physiol Pharmacol 2010; 54(3)
Transcriptomics of Human First Trimester Placental Villi
examination and ultrasonographic evaluation. All women had voluntarily donated their samples after understanding the purpose of the proposed study and gave their written consents. The Ethics Committee of the All India Institute of Medical Sciences and the Ethics Committee of Sitaram Bhartia Institute of Science & Research had approved the research study. Placental samples were collected in sterile ice-cold phosphate buffered saline (PBS, pH 7.4) and transported on ice to the laboratory within 10 min after collection for further processing as described below. Placental villi were isolated from freshly collected placentas and used for RNA extraction as described previously (9). Briefly, tissue samples were quickly washed in icecold phosphate buffered saline to get rid of blood and debris, and separated from surrounding membranes, and villous portions were dissected from decidual and embryonic tissue under a dissecting microscope and further washed with sterile cold phosphate buffered saline (PBS, pH 7.4), and immediately subjected to RNA extraction as described in the following section. Arbitrarily selected small pieces from each sample were processed for chemical fixation and preparation of epoxy blocks for 1 μm section histology as described earlier (10). Table I shows the details of the samples. TABLE I :
Group
1 2 3
Details of groups and tissue samples.
Description Samples (number) (weeks of gestation) Collected Extracted 1 Selected for arrays 2 6 7 8
8 9 7
7 8 7
6 6 6
shows numbers of samples that yielded sufficient villi in the respective groups. 2 based on RNA quality assessment from electrophoresis, yield and RIN score (≥8.0) as described in Materials and Methods. 1
RNA
237
extraction
The methodological details of RNA extraction have been given elsewhere (9). Briefly, total RNA was extracted using Trizol (Agilent Technologies Singapore Pvt. Ltd., Shung Avenue, Singapore) and cleaned up with DNase I (Sigma Chemical Co., St. Louis, Missouri, USA) and subjected to re-extraction when it was necessary. The yield and purity of the extracted RNA was checked using standard protocols of spectrophotometric absorbance at 260 nm, ratio between 260 nm and 280 nm (>1.8) and 1% agarose gel electrophoresis (11). Furthermore, RIN score of individual sample was determined using Agilent 2100 Bioanalyzer, RNA 6000 Nano LabChip kit and Agilent 2100 Expert Software (Agilent Technologies, Inc., Santa Clara, CA, USA) (12). Four samples that could not yield either sufficient amount of RNA or acceptable RIN score (≥8.0) were discarded (Table I). All accessory chemicals were purchased from Sigma Chemical Co. (St. Louis, Missouri, USA) and Invitrogen (Agilent Technologies Singapore Pvt. Ltd., Shung Avenue, Singapore). Nylon
array
cDNA arrays were produced containing custom tailored sequence verified 408 clones spotted (10 ng each in 200 nm) at least in duplicates and UV cross-linked on 6 cm × 6 cm nylon membranes along with the quality assurance by the Clontech (Takara Bio Inc., Shiga, Japan). Details of the gene products examined in the present study are given in Appendix I. Complex probe synthesis and array hybridization
The complex probe synthesis, hybridization and post-hybridization stringency wash were performed as described previously (9). Briefly, radiolabelled cDNA probes were synthesized by labeling 5 μg of total RNA w i t h [ α 33 P ] d A T P ( B h a b h a A t o m i c R e s e a r c h
238
Khan et al
Centre, Hyderabad, India) followed by removal of unincorporated radioactivity on columns using a kit obtained from Clontech (Takara Bio Inc., Shiga, Japan). cDNA probes were denatured and subjected to hybridization at 65°C for 16 h in hybridization buffer at 1 × 106 cpm/ml with the array membrane treated previously with pre-hybridization buffer. The details of hybridization buffer are given elsewhere (13). The process of complex probe synthesis and array hybridization for each sample was performed in duplicate. The hybridized membranes were subjected to step-wise stringency wash in SSC buffer and exposed to low energy storage phosphor screens (BioRad Laboratories, Hercules, CA, USA) for 48 h and scanned at 50 μm resolution in a BioRad Molecular Image Fx Scanner (BioRad Laboratories, Hercules, CA, USA). Data
analysis
Array analysis : Hybridization signals were quantified using ProArray Analysis software (BioRad Laboratories, Hercules, CA, USA). Analysis of data retrieved from separate membranes with same RNA samples yielded QC statistics highly concordant with that of the manufacturer, and it revealed more than 95% confidence level. In order to screen out biological variation from technical variability, integrated signal analysis for a given membrane was done and signal spots that were low after averaging as compared to background values were removed. Data normalization, averaging, calculation of relative abundance of transcripts, ratio analysis and fold changes were performed on log transformed data using Arrayassist (Stratagene, North Torey, La Jolla, CA, USA) and GeneSpring 11 version (Agilent Technologies, Santa Clara, CA, USA). For statistical analysis of data for unraveling the significant (P
