Corrections for mRNA extraction and sample normalization errors find increased mRNA levels may compensate for cancer haplo‐insufficiency

Abstract

The relative mRNA levels of differentially expressed (DE) and housekeeping (HK) genes of six aneuploid cancer lines with large-scale genomic changes identified by SNP/SKY analysis were compared with similar genes in diploid cells. The aneuploid cancer lines had heterogeneous genomic landscapes with subdiploid, diploid, and supradiploid regions and higher overall gene copy numbers compared with diploid cells. The mRNA levels of the haploid, diploid, and triploid HK genes were found to be higher after correction of easily identifiable mRNA measurement errors. Surprisingly, diploid and aneuploid HK gene mRNA levels were the same by standard expression array analyses, despite the higher copy numbers of the cancer cell HK genes. This paradoxical result proved to be due to inaccurate inputs of true intra-cellular mRNAs for analysis. These errors were corrected by analyzing the expression intensities of DE and HK genes in mRNAs extracted from equal cell numbers (50:50) of intact cancer cell and lymphocyte mixtures. Correction for both mRNA extraction/sample normalization errors and total gene copy numbers found the SUIT-2 and PC-3 cell lines' cancer genes both had ∼50% higher mRNA levels per single allele than lymphocyte gene alleles. These increased mRNA levels for single transcribed cancer alleles may restore functional mRNA levels to cancer genes rendered haplo-insufficient by the genetic instability of cancer. © 2013 Wiley Periodicals, Inc.

Variation in transcriptome size: are we getting the message?

“…Transcriptional amplification and gene expression responses to cancer and Rett syndrome Genes that mediate global transcriptional amplification play central roles in cancer Lovén et al 2012;Weaver et al 2014) and Rett syndrome (Li et al 2013), and the assumption that transcriptome size is constant has likely led to misinterpretation of expression data in studies of both diseases. As mentioned above, Myc is a transcription factor that causes global "transcriptional amplification" Lovén et al 2012;Nie et al 2012), though recent work suggests that this effect is indirect (Sabo et al 2014) and that Myc has gene-specific regulatory effects as well (Walz et al 2014).…”

“…Myc expression varies widely among different cancer cell lines ), yet numerous studies profiling gene expression in cancer cells have used standard normalization procedures that assume constant transcriptome size . Additionally, many cancer cells are aneuploid and/or polyploid (e.g., Weaver et al 2014), both of which can cause changes in transcriptome size (Coate and Doyle 2010;Birchler 2014). As a result of global transcriptional amplification induced by Myc and/or alterations in chromosome number, changes in the expression of individual genes (including c-myc, the gene that encodes Myc) on a percell basis will be obscured or exaggerated in transcriptomenormalized expression data (Fig.…”

Identification and characterization of genes with absolute mRNA abundances changes in tumor cells with varied transcriptome sizes

“…Currently, CNAs in cancer genomes are determined by comparing the measurements for the same amounts of DNA extracted from cancer and normal tissues, based on the assumption that the overall yields of DNA per cell (genome sizes) of different cell types are approximately the same. However, this assumption is least likely to hold because many cancers are aneuploid and/or polyploid [3]. The wrong assumption might lead to a serious consequence because CNAs are often used to determine cancer driver genes [47, 48].…”

“…The REOs comparison algorithm cannot be used to detect CNAs because, theoretically, the DNA intensity signals in normal cells should be equal. Due to the same problem of cancer cell aneuploid and/or polyploid [3], DNA methylation analyses using bisulfite-Seq data based on the same amount of DNA are also problematic [9]. Notably, the average beta value of a given locus measured by the Illumina bead-array can be interpreted as an estimate on the proportion of methylated cells to all measured cells [51–53].…”

“…However, this assumption does not hold under many biological conditions. For example, high expression level of c-Myc can induce global transcriptional amplification of cancer cells [2] and many cancer cells are aneuploid and/or polyploid [3], both of which may cause a change in the transcriptome size [4, 5]. Consequently, if two samples being compared are different in transcriptome sizes but still the same amounts of RNA are used which will result in different numbers of cells between the two measured samples, a direct comparison of the measurement values of the two samples can only identify transcripts with changed cellular concentration which might have no changes in absolute mRNA abundances [6].…”