Fig 1: A. Heatmap of gene expression levels between the JH cohort and 319 genes with ASEs. Heatmap and hierarchical clustering of gene expression levels between samples in the Johns Hopkins cohort and 319 genes with alternative splice event numbers in each sample. There was robust correlation between cluster and ASE. As ASE increased, a defined cluster was obtained. 1. B. Representative results of proliferation assay screening. Representative results of proliferation screening assay using the siRNA of 13 candidate genes. Growth curves of BICR22 cells transfected with si-CPSF1 were compared with si-control cells on day 3. P value was calculated using Student’s t-test. *: P < 0.05, **: P < 0.01. These experiments were performed in triplicate in each siRNA. C. Comparison of ASEs between the CPSF1-high and -Normal groups in TCGA. The CPSF1-high group was defined as samples with a Z score > 2 by cBioPortal (http://cbioportal.org). The other samples were defined as the Normal group. Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, outliers are represented by dots. The number of ASEs was significantly different between the Normal- and high-expression CPSF1 groups in TCGA HPV-negative cohort. P value was calculated by using Student’s t-test. **: P < 0.01.
Fig 2: A. Overexpression of CPSF1 promoted cell proliferation. Cell proliferation was detected in SCC17B and SCC090 cell lines after transfection of empty vector and CPSF1 vector. The proliferation of the SCC090 cells overexpressing CPSF1 was significantly increased compared to empty vector cells. P value was calculated using Student’s t-test. *: P < 0.05, **: P < 0.001. These experiments were performed in triplicate. B. Overexpression of CPSF1 induced tumorigenicity. SCC090 cells with the empty vector and overexpression of CPSF1 were injected into nude mice and examined for tumorigenic ability. Tumor progression in the nude mice injected with cells overexpressing CPSF1 was faster than that in the nude mice inoculated with empty-vector-transfected cells. Tumor weight in the CPSF1 group was significantly heavier than in the empty vector group. P value was calculated using Student’s t-test. *: P < 0.05. These experiments were performed in triplicate. C. CPSF1 immunohistochemistry on tissue microarray. C shows the average of CPSF1 expression by IHC staining between non-neoplastic epithelia and squamous cell carcinoma cores from the tissue microarray. Immunohistochemistry of the tissue microarrays demonstrated statistically significant overexpression of CPSF1 in head and neck squamous cell carcinoma tumor as compared to non-neoplastic squamous epithelial tissue (p = 0.016). P value was calculated using Student’s t-test. *: P < 0.05.
Fig 3: A. Colony formation assay. Knockdown of CPSF1 by inducible sh-CPSF1 decreased the colony formation ability of BICR22 cells. Cells containing sh-control and sh-CPSF1 were incubated for 14 days with or without DOX. The number of colonies formed in the sh-CPSF1 cells with DOX was significantly decreased compared with the sh-control cells. These experiments were performed in triplicate. P value was calculated using Student’s t-test. *: P < 0.05, **: P < 0.001. B. Results of the proliferation assay. Cell proliferation was detected in the BICR22 cell lines after stable transfection of inducible sh-control or sh-CPSF1. The proliferation of the cells with sh-CPSF1 and DOX was inhibited compared to other cells. A two way ANOVA was conducted to compare the main effects and interaction between use of DOX and cell line. There as a significant main effect of use of DOX (p<0.001). Also, there was a significant interaction between DOX and cell lines (p<0.01). As the interaction was significant, Post hoc analyses was performed as a multiple comparison of all groups. Post hoc analyses using Steel-Dwass test showed that CPSF1+ DOX group had a significant difference compared to other three groups (p<0.05). These experiments were performed in triplicate. *: P < 0.05. C. Apoptosis assay. Apoptotic cells were measured three times after induction with 1 μg/ml DOX in sh-control- and sh-CPSF1-cells by flow cytometric analysis. Apoptosis in the sh-CPSF1 cells was significantly increased compared with the sh-control cells. P value was calculated using Student’s t-test. *: P < 0.05, **: P < 0.001.
Fig 4: A. Exon Isoform analysis. To validate the junction expression in the knockdown datasets, two candidate junctions (TGFBI (chr5:135390550–135390958) and UBE2C (chr20:44443109–44444493)) in significantly differentially spliced genes were analyzed. TGFB1 variant was novel and UBE2c variant was previously annotated. Each of these alternate splice events is predicted to alter the peptide sequence within the expressed exons: chr5:135390550–135390958 represents accepter intron retention from an early splice accepter site that introduces an early stop codon in TGFB1; chr20:44443109–44444493 introduce cassette exons, which each skips a single canonically expressed exon in UBE2C. B. Validation of junction expression in knockdown dataset by qRT-PCR. To validate the junction expression between conditions in the knockdown datasets, UBE2C (chr20:44443109–44444493) and TGFBI (chr5:135390550–135390958) were selected for RT-PCR. The junction expression of UBE2C and TGFBI in RNA-seq were decreased after knockdown of CPSF1. The change obtained from RNA-Seq was also confirmed by qRT-PCR using same cell line.C. Junction expression between normal and tumor sample with TCGA HPV negative cohorts. Validation using the junction expression data from TCGA HPV negative cohort was performed. A box plot was created to examine the relationship between normal samples and tumor samples in TCGA. In the comparison of junction expression between normal and tumor in TCGA HPV negative cohort, a significant increase of junction expression in tumor was confirmed if this junction expression was decreased in the knockdown cell line.
Supplier Page from OriGene Technologies for CPSF1 Human siRNA Oligo Duplex (Locus ID 29894)