Analysis of gene expression in a complex differentiation hierarchy by global amplification of cDNA from single cells

Many differentiating tissues contain progenitor cells that differ in their commitment states but cannot be readily distinguished or segregated. Molecular analysis is therefore restricted to mixed populations or cell lines which may also be heterogeneous, and the critical differences in gene expression that might determine divergent development are obscured. In this study, we combined global amplification of mRNA transcripts in single cells with identification of the developmental potential of processed cells on the basis of the fates of their sibling cells from clonal starts.

We analyzed clones of from four to eight hemopoietic precursor cells which had a variety of differentiative potentials; sibling cells generally each formed clones of identical composition in secondary culture. Globally amplified cDNA was prepared from individual precursors whose developmental potential was identified by tracking sibling fates. Further cDNA samples were prepared from terminally maturing, homogeneous hemopoietic cell populations. Together, the samples represented 16 positions in the hemopoietic developmental hierarchy. Expression patterns in the sample set were determined for 29 genes known to be involved in hemopoietic cell growth, differentiation or function. The cDNAs from a bipotent erythroid/megakaryocyte precursor and a bipotent neutrophil/macrophage precursor were subtractively hybridized, yielding numerous differentially expressed cDNA clones. Hybridization of such clones to the entire precursor sample set identified transcripts with consistent patterns of differential expression in the precursor hierarchy.

Tracking of sibling fates reliably identifies the differentiative potential of a single cell taken for PCR analysis, and demonstrates the existence of a variety of distinct and stable states of differentiative commitment. Global amplification of cDNA from single precursor cells, identified by sibling fates, yields a true representation of lineage- and stage-specific gene expression, as confirmed by hybridization to a broad panel of probes. The results provide the first expression mapping of these genes that distinguishes between progenitors in different commitment states, generate new insights and predictions relevant to mechanism, and introduce a powerful set of tools for unravelling the genetic basis of lineage divergence.