Quantification of gene expression in single cells

University dissertation from Department of Clinical Sciences, Lund University

Abstract: Studies of apparently homogeneous cell populations and single cells often give highly divergent results. Cells exhibit varying responsiveness to stimuli and gene expression and they are in many aspects stochastic and unpredictable. We have developed a method to measure gene expression quantitatively in individual cells with real-time RT-PCR.

mRNA for hormones, ion channels and enzymes in pancreatic alpha- and beta-cells were quantified. The distribution of transcript levels were highly skewed and was best described by a lognormal distribution. Thus, the geometric--and not the commonly used arithmetic--mean value is the appropriate measure of average expression level. In beta-cells, insulin mRNA levels were increased in response to glucose stimulation; an effect due to an increased fraction of cells with high expression. The insulin genes Ins1 and Ins2 have similar promoter regions and were indeed co regulated within single beta-cells.

Na-channels in alpha- and beta-cells display very different inactivation properties (being separated by 40 mV). We measured hormone mRNA and all Na-channel isoforms in single cells and correlated gene expression with patch-clamp recordings. Cell type-specific expression of Na-channel isoforms can partly explain the divergent inactivation.

Early differentiation of human embryonic stem cells involves the transcription factors Pou5f1, Nanog and Sox2. We quantified their expression in single stem cells and observed that they are not correlated with each other. Instead Pou5f1 correlates with the transcription factors Id1 and Id3.

We conclude that quantitative gene expression measurements on single cells allow: 1) studies of cell population heterogeneity and noise in gene expression; 2) exploration of genes that are co-regulated; and 3) correlation of gene expression with functional properties such as electrophysiological properties.