Unknown,Transcriptomics,Genomics,Proteomics

Dataset Information

The effect of freeze-thaw cycles on gene expression levels in lymphoblastoid cell lines

ABSTRACT: Epstein-Barr virus (EBV) transformed lymphoblastoid cell lines (LCLs) are a widely used renewable resource for functional genomic studies in humans. The ability to accumulate multidimensional data pertaining to the same individual cell lines, from complete genomic sequences to detailed gene regulatory profiles, further enhances the utility of LCLs as a model system. However, the extent to which LCLs are a faithful model system is relatively unknown. We have previously shown that gene expression profiles of newly established LCLs maintain a strong individual component. Here, we extend our study to investigate the effect of freeze-thaw cycles on gene expression patterns in mature LCLs, especially in the context of inter-individual variation in gene regulation. We found a profound difference in the gene expression profiles of newly established and mature LCLs. Once newly established LCLs undergo a freeze-thaw cycle, the individual specific gene expression signatures become much less pronounced as the gene regulatory programs in LCLs from different individuals converge to a more uniform profile, which reflects a mature transformed B cell phenotype. As expected, previously identified eQTLs are enriched among the relatively few genes whose regulations in mature LCLs maintain marked individual signatures. We thus conclude that findings and insight drawn from gene regulatory studies in mature LCLs are generally not affected by artificial nature of the LCL model system and are likely to faithfully reflect regulatory interactions in primary tissues. However, our data indicate that many aspects of primary B cell biology cannot be observed and studied in mature LCL cultures. We obtained unpurified buffy coats (of a unit of blood) from six unrelated healthy individuals of Caucasian ethnicity (age range: 20-45). We isolated CD20+ B cells and established six independent cultures of LCLs between October 2009 and January 2010. From each of these samples, we obtained genome wide gene expression data using Illumina HumanHT-12 v3 Expression BeadChip arrays. We refer to data from these experiments as M-bM-^@M-^\cycle 0M-bM-^@M-^] to acknowledge the fact that the LCLs from this cycle were newly established and therefore not frozen/thawed. Between February 2011 and October 2012, we thawed each of these LCL cultures every 3 months and cultured them until obtaining ~10 million cells (February 2011=cycle 1, June 2011=cycle 2, October 2011=cycle 3, February 2012=cycle 4, June 2012=cycle 5, and October 2012=cycle 6). We used Illumina HumanHT-12 v4 Expression BeadChip arrays and obtained genome wide gene expression data on cycle 2, cycle 4 and cycle 6 LCLs. At each hybridization batch we included a subset of RNA samples that were also hybridized in a previous batch. This allowed us to effectively consider the batch effect on gene expression profiles. Our final dataset included genome wide gene expression data from 187 samples collected at 4 different time points.

ORGANISM(S): Homo sapiens

SUBMITTER: Minal Caliskan

PROVIDER: E-GEOD-58942 | biostudies-arrayexpress |

REPOSITORIES: biostudies-arrayexpress

ACCESS DATA

Similar Datasets

Project description:Renewable in vitro cell cultures, such as lymphoblastoid cell lines (LCLs), have facilitated studies that contributed to our understanding of genetic influence on human traits. However, the degree to which cell lines faithfully maintain differences in donor-specific phenotypes is still debated. We have previously reported that standard cell line maintenance practice results in a loss of donor-specific gene expression signatures in LCLs. An alternative to the LCL model is the induced pluripotent stem cell (iPSC) system, which carries the potential to model tissue-specific physiology through the use of differentiation protocols. Still, existing LCL banks represent an important source of starting material for iPSC generation, and it is possible that the disruptions in gene regulation associated with long-term LCL maintenance could persist through the reprogramming process. To address this concern, we studied the effect of reprogramming mature LCLs to iPSCs on the ensuing gene expression patterns within and between six unrelated donor individuals. We show that the reprogramming process results in a recovery of donor-specific gene regulatory signatures. Since environmental contributions are unlikely to be a source of individual variation in our system of highly passaged cultured cell lines, our observations suggest that the effect of genotype on gene regulation is more pronounced in the iPSCs than in the LCL precursors. Our observations indicate that iPSCs can be a powerful model system for studies of phenotypic variation across individuals in general, and the genetic association with variation in gene regulation in particular. We further conclude that LCLs are an appropriate starting material for iPSC generation. Whole genome gene expression data was collected for 34 samples including 17 iPSC lines and 17 LCL lines on the Illumina HT-12 v4 Expression BeadChip array platform. Three biological replicates of each individual were included in the study, except for one individual for which only two replicates were obtained.

Project description:DNA from Epstein-Barr virus (EBV)-transformed lymphocyte cell lines (LCLs) has proven very useful for studies of genetic sequence polymorphisms. Whether EBV-LCL DNA is suitable for methylation studies is less clear. We conduct a genome-wide methylation investigation using an array set with 45 million probes to investigate the methylome of EBV-LCL DNA and technical duplicates of whole blood (WB) DNA from the same 10 individuals. Methylation sites that show variation between individuals are potentially useful as biomarkers in disease studies. Our comparison is, therefore, focused specifically on these methylation variable sites. The sample correlations (i.e., a measure of whether the rank of the signals remains consistent between two samples from the same individual) for the methylation variable probes ranged from 0.69-0.78 for the WB duplicates and from 0.27-0.72 for WB versus EBV-LCL. To compare the pattern of the methylation signals, we grouped adjacent probes based on their inter-correlations. These analyses showed ~29,000 blocks in WB and ~14,000 blocks in EBV-LCL. Furthermore, merely 31% of the methylated regions detected in WB were detectable in EBV-LCLs. Our study shows that there are substantial differences in the DNA methylation patterns between EBV-LCL and WB. Thus, EBV-LCL DNA cannot be used as a proxy for WB DNA in methylation studies. The methylation of technical duplicates of DNA extracted from whole blood from 10 individuals, as well as DNA extracted from EBV-transformed lymphocyte cell lines (EBV-LCL) from the same samples of whole blood, are investigated using the Affymetrix GeneChip Human Tiling 2.0R array set. In total, 30 samples from 10 individuals were investigated on 7 arrays. The supplementary "GSE35204_ChrXX_bgc_qt.txt" files contain the background-corrected and quantile-normalized data for all 30 samples per chromosome. The files include 32 columns: 'Seq' refers to the chromosome, 'Pos' indicates the position in the genome on that chromosome, and columns 3-32 indicate the sample numbers.

Project description:Purpose: There exists a rich bio-resource of numerous lymphoblastoid cell line (LCL) repositories generated from a wide array of patients, many of them with extensive genotypic and phenotypic data already generated. We have developed a highly efficient LCL to induced pluripotent stem cells (iPSC) reprogramming method and performed whole genome mRNA and miRNA analysis to understand mechanistic changes that take place at the transcriptome and cellular functional level during reprogramming of LCLs into iPSCs and further differentiation. Methods: Applying our optimized protocol which utilizes episomal plasmids encoding pluripotency transcription factors and mouse p53DD - p53 carboxy-terminal dominant-negative fragment and commercially available reprogramming media, we reprogrammed six LCLs into iPSCs and then differentiated them into neural stem cells (NSC). The LCLs, their reprogrammed iPSCs and differentiated NSC (n=18) were sequenced for mRNA and smallRNA on an illumina HiSeq 2500. Differential gene expression analysis was performed between LCL-iPSC and iPSC-NSC pairs in combination with functional annotations and Ingenuity® Pathway Analysis (IPA). Results: Our LCL reprogrammed iPSCs express the majority of genes and miRNAs known to contribute to stemness in human ESCs. The functional enrichment analysis of the up-regulated genes and activation of human pluripotency pathways in the reprogrammed iPSCs showed that the generated iPSCs have a transcriptional and functional profile very similar to that of human ESCs. The reprogrammed iPSCs also showed the potential to differentiate into cells of all three germ layers. Significantly, the transcriptomic effect of EBV encoded oncoproteins which were very pronounced in LCLs, were significantly inhibited in reprogrammed iPSCs. The transcriptomic and functional enrichment analysis of the NSC differentiated from the reprogrammed iPSCs showed that they share a functional profile of self-renewing NSCs. Conclusions: We have been able to develop a MEF feeder free protocol for efficient and reproducible reprogramming of LCLs into iPSC. In addition our comprehensive analysis of genome wide miRNA and mRNA of LCLs, their reprogrammed iPSC and differentiated NSCs provides important documentation of differentially expressed genes and miRNAs and their functional consequences during LCL to iPSC reprogramming and NSC differentiations that were previously unknown.

Project description:Transcript abundance results from the balance between transcription and mRNA decay, and varies pervasively in humans. We have examined the effect of DNA variation on mRNA half-life differences by conducting a genome-wide survey of mRNA stability in seven human HapMap lymphoblastoid cell lines (LCLs). We determined the mRNA half-life for each gene from the ratio of 4-thio-uridine (4sU)-labeled nascent RNAs to total RNAs. 5,145 (46%) of 11,132 analyzed genes showed inter-individual mRNA half-life differences at a false discovery rate, FDR<0.05. As previously reported, we found transcription to be the main factor influencing transcript abundance. Although mRNA half-life explained only ~6% of transcript abundance on average, it explained ~16% for the subset of genes (~10%) showing inter-individual mRNA half-life differences (P<0.001). We confirmed previously reported correlations of mRNA half-life with transcript length, 3M-bM-^@M-^Y-UTR length, and number of exon-junctions per kb of transcript. The number of miRNA targets in 3M-bM-^@M-^Y-UTRs was negatively correlated with half-life (P=2.2M-CM-^W10-16), a new observation that is consistent with the role of miRNA in inducing mRNA degradation. Notably, coding GC and GC3 content showed positive correlations with mRNA half-life in genes with inter-individual mRNA half-life differences, implying a role of mRNA stability in shaping synonymous codon usage bias. Consistently, G or C alleles of coding SNPs were found associated with longer mRNA half-life (P=0.021). As expected, we also found that nonsense SNPs were associated with shorter mRNA half-life (P=0.009). Our results strongly suggest that inter-individual mRNA stability differences are widespread and affected by DNA sequence and composition variation. A total of 7 HapMap LCLs were used to measure mRNA half-life. Total RNAs and the 4sU-labeled-newly synthesized RNAs (nascent RNAs) were isolated from the same cell culture and were assayed simultaneously with human Exon array. For 3 LCLs, we included 3 biological replicates (i.e., independent cell cultures) and for 1 LCL we also included technical duplicates. mRNA half-life was calculated from the ratio of nascent RNAs/total RNAs. We used ANOVA to test inter-individual difference of mRNA half-life between 3 subjects who have biological replicates and technical duplicates. We examined the Spearman rank correlation of mRNA half-life with a number of gene features, including transcript length, intron length, 5'-UTR length and folding energy, 3'-UTR length and folding energy, microRNA target sites, GC and GC3 contents, etc. We also performed linear regression to test the effects of specific type of sequence variants (nonsense SNPs, SNPs within miRNA target sites, and coding synonymous and nonsynonymous SNPs) on mRNA half-life across 3 subjects that have whole genome sequencing data available (1000 genome project June 2011 release).

Dataset Information

The effect of freeze-thaw cycles on gene expression levels in lymphoblastoid cell lines

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets