Screening small molecules and drugs for activity to modulate alternative splicing, we found that amiloride, distinct from four other intracellular pH-affecting analogues, could normalize the splicing of BCL-X, HIPK3 and RON/MISTR1 transcripts in human hepatocellular carcinoma Huh-7 cells. To elucidate the underlying mechanisms, our proteomic analyses of amiloride-treated cells detected hypo-phosphorylation of splicing factor SF2/ASF and also decreased levels of SRp20 and two un-identified SR proteins. We further observed decreased phosphorylation of AKT, ERK1/2 and PP1, while increased phosphorylation of p38 and JNK, suggesting that amiloride treatment down-regulated kinases and up-regulated phosphatases in the signal pathways known to affect the splicing factor protein phosphorylation. The amiloride effects of splicing factor protein hypo-phosphorylation andnormalizedoncogenic RNA splicing were both abrogated by pre-treatment with a PP1 inhibitor. We then performed global exon array analysis of Huh-7 cells treated with amiloride for 24 hours. Using gene array chips (Affymetrix GeneChip Human Exon 1.0 ST Array of >518000 exons of 42974 genes) for exon array analysis (set parameters of correlation coefficient 0.7, splicing index -1.585 , and log2 ratio -1.585), we found that amiloride influenced the splicing patterns of 551 genes involving at least 584 exons, which included 495 known protein-coding genes involving 526 exons, many of which play key roles in functional networks of ion transport, extracellular matrix, cytoskeletons and genome maintenance. Cellular functional analyses revealed subsequent invasion and migration defects, cell cycle disruption, cytokinesis impairment, and lethal DNA degradation in amiloride-treated Huh-7 cells. This study thus provides mechanistic underpinnings for exploiting small molecule modulation of abnormal RNA splicing for cancer therapeutics.
Small molecule amiloride modulates oncogenic RNA alternative splicing to devitalize human cancer cells.
Cell line
View SamplesAlternative splicing is a mechanism for increasing the protein variety of a limited number of genes. Studies have shown that aberrant regulations of the alternative splicing of apoptotic gene transcripts may contribute to the development of cancer. In this study, we isolated 4ß-Hydroxywithanolide E (4bHWE) from the traditional herb Physalis peruviana, and analyzed its biological effects in cancer cells. The results demonstrated that 4bHWE modulates the alternative splicing of apoptotic genes (e.g., HIPK3, SMAC/DIABLO, and SURVIVIN), changes the expression level of splicing factors (e.g., hnRNP C1/C2, ASF/SF2, SRp20, and SRp55), and induces histone tail posttranslational modifications (e.g., H3K27me1, H3K27me2, H3K36me3, and H3K79me1). Pretreatment with okadaic acid to inhibit protein phosphatase-1 could partly relieve the effects of 4bHWE on the alternative splicing of HIPK3 and SMAC/DIABLO transcripts, as well as on the dephosphorylation of ASF/SF2. Genome-wide detection of alternative splicing further indicated that several other apoptosis-related genes are also regulated by 4bHWE, including APAF1, CARP-1, and RIPK1. Moreover, we extended our study to apoptosis-associated molecules, detecting an increasing level of CASPASE-3 activity and cleavage of poly ADP-ribose polymerase in 4bHWE-induced apoptosis. Furthermore, in vivo experiments showed that the treatment of tumor-bearing mice with 4bHWE resulted in a marked decrease of tumor size and weight. Taken together, this study is the first to show that 4bHWE affects alternative splicing through the modulations of splicing factors, providing a novel view of the antitumor mechanism of 4bHWE. Overall design: Examination of the global genes with altered alternative splicing in 4bHWE-treated Huh-7 cells.
4β-Hydroxywithanolide E Modulates Alternative Splicing of Apoptotic Genes in Human Hepatocellular Carcinoma Huh-7 Cells.
Specimen part, Treatment, Subject
View SamplesWe demonstrate that GLUT4 up-regulation significantly increased cell migration and invasion in lower magligance head and neck cancer cell lines in vitro.
Glucose transporter 4 promotes head and neck squamous cell carcinoma metastasis through the TRIM24-DDX58 axis.
Specimen part, Cell line
View SamplesPurpose: The cause of labour initiation has yet to be fully elucidated for human pregnancy. This has hindered attempts to find effective therapies for the prevention of preterm labour, which affects up to 10% of pregnancies in the UK and it is the most dominant cause of perinatal death (75% of all cases). The myometrium of the uterus is where contractions that characterise labour take place, and it is here where changes at the molecular level responsible for triggering labour potential originate from. We used RNA-Seq-based mRNA sequencing technology in an attempt to identify mRNA transcripts that are differentially expressed in the myometrium upon the onset of labour, by comparing the expression profiles of tissues samples that represented non-labour (TNL), early labour (TEaL, = 3 cm cervical dilation) and established labour (TEsL, > 3 cm cervical dilation) states at term (> 37 weeks gestation) pregnancy. Methods: Myometrial biopsies from women undergoing Caesarean section were collected in accordance with the Declaration of Helsinki guidelines, and with approval from the local research ethics committee for Chelsea & Westminster Hospital (London, UK; Ethics No. 10/H0801/45). Informed written consent was obtained from all women who participated. Biopsies were excised from the upper margin of the incision made in the lower segment of the uterus, immediately washed with Dulbecco's phosphate-buffered saline (Sigma) and dissected into pieces approximately measuring 2-3 mm3. For RNA study, biopsies were immersed in RNAlater (Sigma) within 6 minutes after biopsy excision from the uterus and stored at 4°c overnight, before being taken out of RNAlater solution to be frozen for long-term storage at -80°c. For CHIP study, biopsies were snap-frozen in liquid nitrogen and stored at -80°c. All specimens were categorised into four groups according to their labour stages: preterm not in labour (PTNL, n=6), term not in labour (TNL, n=8), term in early labour (TEaL, defined as cervical dilatation <3 cm, n=8) and term in established labour (TEsL, defined as cervical dilatation >3 cm, n=6). For each sample, 60-100 mg of myometrium tissue were extracted in TRIzol (Life Technologies) by mechanical homogenisation in a Precellys 24 bead-based homogeniser using 5 cycles of 5000 rpm for 20 seconds, before chloroform treatment and centrifugation at 4°c. RNA was extracted from the aqueous phase of centrifuged homogenates using the TRIzol Plus RNA Purification kit (Life Technologies) with on-column DNase treatment prior to elution, all according to manufacturer's instructions. Final RNA samples were stored at -80°c. The quantity and quality RNA was measured using a Nanodrop ND-1000 spectrophotometer (LabTech), Qubit fluorimeter (Life Technologies) and Bioanalyser 2100 (Agilent Technologies). Preparation of cDNA libraries was carried out using the TruSeq Stranded mRNA Sample Preparation kit (Illumina), following the high-throughput sample (HT) protocol. The quantity and quality of cDNA libraries were also tested by a Qubit fluorimeter and Bioanalyser 2100. TruSeq Stranded libraries were then multiplexed and sequenced with the average of 42 million DNA fragments per sample (100 bp paired-end reads). Quality control was performed using FastQC software (version 0.11.2). RNA-Seq reads was aligned to the GRCh38 Homo sapiens reference genome downloaded from Ensembl (release 81) with HISAT version 2.0.1 using parameters of --known-splicesite-infile --dta-cufflinks --rna-strandness RF --phred33 –p 4 -q. A list of known splice sites generated from the Ensembl GTF file using an accessory python script included in the HISAT2 package was provided to --known-splicesite-infile, of which HISAT2 will make use to assist the alignment of reads spanning two or more exons. Ensembl annotated a total of 65,989 genes, which includes 20,276 protein coding genes. As one human gene usually contains multiple transcript models, we thus conducted a transcript merging procedure to produce gene level models for expression analysis. Specifically, exons labeled as 'retained_intron' were first excluded, then overlapping interval exons of each gene were merged and a final gene level model was produced in GFF format. Only uniquely mapped (i.e. reads with the tag of NH:i:1) reads were used to produce gene read counts and calculate gene expression levels. The raw read count matrix was normalised with DESeq2 (version 1.6.3). Expression level of each gene in each sample was represented as RPKM (reads per kilobase per million mapped reads). Differentially expressed genes (DEGs) between two groups of samples were identified with DESeq2 (version 1.6.3), edgeR (version 3.8.6) and Cuffdiff (version 2.2.1). For DESeq2 and edgeR, we used the normalised read count matrix as input, and for Cuffdiff, we used the alignment bam files with uniquely mapped reads as input. Raw p-values were adjusted by FDR to produce q-values, and q-value of 0.05 were chosen as the cut-off for statistical significance in DESeq2, edgeR as well as Cuffdiff. Results: 22 RNA samples from three different labour groups were sequenced and an average of 53 million reads were obtained from each sample. More than 97.34% of reads were successfully aligned to GRCh38 reference human genome and the unique concordant pair ratio was greater than 92.39%. In total, 60,593 genes were mapped with the following criteria: (1) at least one RNA-seq read assigned to a gene; (2) we only assign a read to a gene when > 90% of this read falls into the exon regions of this gene. The principal component analysis (PCA) of these 22 samples showed that TNL and TEsL samples formed two distinct clusters whereas the TEaL group featured relatively great internal differences. Nevertheless, half of the samples from TEaL group was clustered with the TNL group and the other half was more separated yet closer to two samples of TEsL group. To determine the transcripts associated with labour, three software packages (DESeq2, EdgeR and Cuffdiff) were used to perform differential expression The transcript with a q value <0.05 in at least two methods was defined as a shared differentially expressed gene (DEG). As a result, 132 and 399 genes were identified comparing TNL with TEaL and TEsL, respectively, whereas no gene was significantly differentially expressed between TEaL and TEsL groups. Due to big differences among individual samples, in this study, the expression fold change (FC) was calculated as the ratio of median reads per kilobase per million mapped reads (RPKMs) with the bigger median RPKM divided by the smaller median RPKM. In order to minimise the noise derived from genes with low expression but high FC, we further filtered gene lists according to the following rational: if the original value of any median RPKM was <1, we artificially turned it into 1 before calculating the FC. Finally, two robust gene lists with an expression FC >1.5 between two groups (TNL vs. TEaL and TNL vs. TEsL) were generated containing 70 and 232 DEGs, respectively. Conclusions: This study, for the first time, identifies differentially expressed genes (DEGs) and pathways that are involved in the myometrial transition from non-labouring to labouring phenotype by using samples from different stages of pregnancy and labour. The DEG lists are generated by subjecting the raw data to three sofware packages (DESeq2, edgeR and Cuffdiff), which makes the yield DEGs more robust. We conclude that some early responsive genes in circadian clock and inflammation pathways are likely to account for the labour onset and no significant changes are found on the transcription level once the labour starts. Overall design: Comparisons made between no labour (TNL, n = 8), early labour (TEaL, n = 8) and established labour (TEsL, n = 6) lower segment myometrial tissue samples, which were collected during Caesarean section (CS) with informed written consent
Myometrial Transcriptional Signatures of Human Parturition.
Subject
View SamplesThe orthotopic transplantation of human OEC-M1 cells in immune-compromised mice was established to feasibly study tumorigenesis and lymph node metastasis of OSCC.
Insulin-like growth factor-independent insulin-like growth factor binding protein 3 promotes cell migration and lymph node metastasis of oral squamous cell carcinoma cells by requirement of integrin β1.
Specimen part, Cell line
View SamplesTo exmaine the PTHLH stimulated genes in Ca9-22 cells, we preformed the Affymetrix Human Genome U133 Plus 2.0 Array with empty vector or PTHLH expression vector. The raw data were normalized by GeneSpring GX software and up-load with raw values.
Parathyroid Hormone-Like Hormone is a Poor Prognosis Marker of Head and Neck Cancer and Promotes Cell Growth via RUNX2 Regulation.
Cell line
View SamplesThe early-onset breast cancer patients (age 40) often display higher incidence of axillary lymph node metastasis, and poorer five-year survival than the late-onset patients. To identify the genes and molecules associated with poor prognosis of early-onset breast cancer, we examined gene expression profiles from paired breast normal/ tumor tissues, and coupled with Gene Ontology and public data base analysis. Our data showed that the expression of GAS7b gene was lower in the early-onset breast cancer patients as compared to the elder patients. We found that GAS7 was associated with CYFIP1 and WAVE2 complex to suppress breast cancer metastasis via blocking CYFIP1 and Rac1 protein interaction, actin polymerization, and 1-integrin/FAK/Src signaling. We further demonstrated that p53 directly regulated GAS7 gene expression, which was inversely correlated with p53 mutations in breast cancer specimens. Our study uncover a novel regulatory mechanism of p53 in early-onset breast cancer progression through GAS7-CYFIP1 mediated signaling pathways.
Wild-type p53 upregulates an early onset breast cancer-associated gene GAS7 to suppress metastasis via GAS7-CYFIP1-mediated signaling pathway.
Sex, Age, Specimen part, Subject
View SamplesTo examine the transcription targets of RUNX2 in OSCC cells, we preformed the Affymetrix Human Genome U133 Plus 2.0 Array with ectopic RUNX2 or empty vectors in Ca9-22 cells.
Dysregulation of RUNX2/Activin-A Axis upon miR-376c Downregulation Promotes Lymph Node Metastasis in Head and Neck Squamous Cell Carcinoma.
Specimen part, Cell line
View SamplesWe have developed a protocol to generate hematopoietic and cardiac derivatives in vitro by Mesp1 induction in ES cells. The goal of this study is to analyze the heterogeneity of Mesp1+ mesoderm by single-cell RNA-seq Overall design: 48 Mesp1-induced single cells were captured using the Fluidigm C1 microfluidic system. RNA extraction, RT and cDNA amplication were then performed according to the manufacturer's manual.
Heterogeneity of Mesp1+ mesoderm revealed by single-cell RNA-seq.
Specimen part, Cell line, Treatment, Subject
View SamplesWe report the expression profiles of MCF10A cells encapsulated in hydrogels of varying stiffness and composition. Cells were encapsulated for 7 days in either 1.) soft alginate and reconstituted basement membrane (rBM), 2.) stiff alginate and rBM, 3,) soft col-1 and rBM, or 4.) stiff col-1. We find global gene expression changes in response to enhanced ECM stiffness, independent of expression changes in response to col-1 exposure. These results provide a comprehensive study of the gene expression changes associated with increased ECM stiffness in addition to the gene expression changes associated with increased col-1 concentration in combination with, and independent of, ECM stiffness. Overall design: Expression profiling of MCF10A cells in four hydrogel conditions were sequenced in duplicate via Illumina HiSeq.
YAP-independent mechanotransduction drives breast cancer progression.
Specimen part, Cell line, Subject
View Samples