Sadhan Majumder sought a better understanding of childhood brain cancer. He ended up finding a new regulator of self-renewal in embryonic stem cells along with a previously unknown mechanism of how this state is maintained.1

Sadhan Majumder

While working on one of the most malignant childhood brain tumours, medulloblastoma, Majumder and his group thought much of the blame for the cancer might lie on a protein called REST (which has the burdensome full name of repressor element-1 silencing transcription factor/neuron-restrictive silencing factor). REST was originally believed to repress the final stages of neuronal differentiation. “We were working on REST in neurogenesis like everyone else,” Majumder recalls, “and we found that in medulloblastoma, abnormal expression of REST maintains the proliferative state of neural stem or progenitor cells and blocks their differentiation. Basically, it maintains the 'stemness' of these cells.” In fact, if REST is introduced artificially into neural stem cells, he says, they will not differentiate. “That told us that it may have a role in self-renewal,” says Majumder. “Then this paper from Gail Mandel came out that showed REST was expressed in embryonic stem cells.”2 He laughs as he described how he decided on his next experiments. “I have a tremendous grasp of the obvious.”

Majumder had worked on how genes are expressed in one- and two-cell mouse embryos as a postdoc, but he moved to neural cells after becoming a professor at the University of Texas M.D. Anderson Cancer Center in Houston. There, along with his lab members, he taught himself to work on embryonic stem (ES) cells and found that REST expression correlates with self-renewal and pluripotency in mouse ES cells. Using two lines of mouse ES cells for each experiment, the team explored this relationship, manipulating REST levels both genetically and with short interfering RNA. Higher levels of REST boosted self-renewal; cells with less REST showed more differentiation. The researchers also identified REST expression in mouse blastocysts. In one effort to find out what REST was doing, the researchers looked for differences in the gene-regulatory molecules known as microRNAs (miRNAs), comparing mouse ES cells that had one versus two functional copies of the Rest gene.

They found several differences, including the expression of miRNAs shown or predicted to be targets of REST. Some were upregulated in cells with only one copy of the Rest gene; others were downregulated, indicating that control networks include double negative-feedback loops. After using databases to identify miRNAs that are likely to target self-renewal genes, the researchers found that REST was bound on chromatin coding for some of these miRNAs and that knocking out REST increased levels of those miRNAs. They next transfected cells with miRNAs identified in their studies; miR-21 had a particularly strong effect, reducing self-renewing capacity by 60% compared with a nontargeting RNA molecule used as a control.

Expression of key pluripotency genes Oct4, Nanog, Sox2 and Myc also decreased when this miRNA was present. The stem cell field has seen a spate of papers on the role of miRNAs in stem cells; last month, one of four genes known to reprogram differentiated cells to an embryonic-like state was found to regulate an miRNA called let-7. Its function, however, is unknown. miR-21 is the first of these gene-regulating molecules that has been shown to affect self-renewal experimentally. Majumder says there's no known connection between these two miRNAs, “but it might suggest that there's a series of microRNA interactions that may have a role in renewal and pluripotency.”

Although the evidence is strong that REST maintains self-renewal by suppressing miRNAs, it is still unclear how miR-21 suppresses self-renewal. Bioinformatics predictions show that miR-21-binding sites are present in transcripts of Sox2 and Nanog but not Oct4 and Myc. The proteins Oct4, Nanog and Sox2 seem to bind to DNA sequences in such a way to enhance REST expression, and REST itself forms part of the Nanog and Oct4 complex. Majumder said that although there's always a temptation to point to one or another protein as the master regulator of renewal, this work shows yet again that self-renewal and pluripotency depend on a highly interconnected network. “There cannot be one single molecule. ES self-renewal is controlled by a circuit, and REST is the newly discovered member of the circuit.”