Supplementary Information
Promoter hypomethylation of EpCAM-regulated bone morphogenetic protein genes in advanced endometrial cancer
Ya-Ting Hsu, Fei Gu, Yi-Wen Huang, Joseph Liu, Jianhua Ruan, Rui-Lan Huang,
Chiou-Miin Wang, Chun-Liang Chen, Rohit R. Jadhav, Hung-Cheng Lai, David G.
Mutch, Paul J. Goodfellow, Ian M. Thompson, Nameer B. Kirma, and Tim Hui-Ming
Huang Tables of contents
Page Table of contents 2
Supplementary Methods 4
Supplementary Figure S1. Summarized sequencing reads and coverage of MBDCap-seq 8
Supplementary Figure S2. Reproducibility test of MBDCap-seq 10
Supplementary Figure S3. Validation of MBDCap-seq by MassARRAY analysis 11
Supplementary Figure S4. Distribution of differentially methylated regions (DMRs) in endometrial tumors relative to normal control 12
Supplementary Figure S5. Network analysis of differential methylation loci by using Steiner-tree analysis 13 Supplementary Figure S6. DNA methylation distribution in early and late stage of the
TCGA endometrial cancer cohort 14
Supplementary Figure S7. Relative expression of BMP genes with EGF treatment in the presence or absence of PI3K/AKT and Raf (MAPK) inhibitors in endometrial cancer cells 15
Supplementary Figure S8. Induction of invasion by EGF in AN3CA and HEC1A cell lines 16
Supplementary Figure S9. Knockdown expression of BMP4 and BMP7 in RL95-2 cells 17
Supplementary Figure S10. Relative expression of BMPs and BMPRs in normal endometrial cell and endometrial cancer cell lines 18
Supplementary Figure S11. Microfluidics-based PCR analysis of EMT gene panel in
RL95-2 cells with or without EGF treatment 19
Supplementary Figure S12. Knockdown expression of EpCAM by different shRNA sequences in RL95-2 cells 20
Supplementary Figure S13. Proposed model for EpICD mediated regulation of BMP genes by EGF 21
2
Supplementary Table 1. Characteristics of endometrial cancer patients 22
Supplemental Table 2. Summary of sequencing reads in endometrial tumors an normal controls 23
Supplementary Table 3. Differentially methylated promoter CpG islands in recurrent relative to non-recurrent tumors 26
Supplementary Table 4. List of primer sequences 55
Supplementary Table 5. List of siRNA and shRNA sequences 59
Supplementary References 60
3
Supplementary Methods
MassARRAY analysis
To quantify methylation levels of CpG islands in clinical samples, the high-throughput
MassARRAY platform (Sequenom) was carried out as described previously (Huang et al. 2009).
Briefly, DNA was bisulfite-converted and subjected to PCR and base-specific cleavage. The cleaved DNA was analyzed by MALDI-TOF mass spectrometry. Methylation data of HAAO,
RXFP3, and miR-129-2 CpG islands were published previously (Huang et al. 2009, 2010).
Definition of methylation types
All known transcripts were divided into 1) promoter CpG island located within 2-kb up- or downstream of its TSS and 2) promoter non-CpG island located within 2-kb windows of its TSS.
Other regions were defined as intragenic CpG island located after 2-kb downstream of a TSS and before 3’-UTR and intergenic CpG island located before >2-kb upstream of a TSS or after
3’-UTR.
Bioinformatics analysis
Sequence reads mapped to unique genome locations were normalized to the sum of unique reads and other reads (e.g., repeat sequences). Unique reads were collected after normalization and directionally extended to 200-bp. The methylation level of each bin size was calculated by accumulating the reads number if the read was located in the bin. Linear normalization method based on the unique mapped read number was used to normalize methylation enrichment: