Cell proliferation and cell death might seem like opposing functions, but oncogenes such as E1A and MYC are able to initiate both. Apoptosis is thought to be a safety mechanism, and is induced when the oncogenic signal to proliferate is recognized as excessive, and hence likely to cause tumorigenesis. The mechanisms by which oncogenes induce cell death have not been clearly established. Zaher Nahle, Scott Lowe and colleagues have investigated this phenomenon using E1A, and have found that the E2F transcription factor, which promotes replication and hence proliferation, has a key role in coordinating these processes.

The authors first investigated whether overexpression of E1A in mouse embryo fibroblasts (MEFs) and normal diploid human fibroblasts (IMR90 cells) affected the protein levels of caspases — the effectors of cell death — and found that both initiator and effector caspases were upregulated by 5–15-fold. A similar increase was observed in cells deficient for either ARF or p53, so E1A must upregulate caspases through a p53-independent pathway.

One of the key targets of E1A is retinoblastoma (RB), and E1A mutants that are unable to inactivate RB do not upregulate caspases. Similarly, RB−/− MEFs expressed higher levels of capases than wild-type cells. Introduction of wild-type RB, but not a tumour-derived mutant that can not bind E2F, into RB-deficient cells represses this caspase expression, and implicates the E2F family of proteins in this apoptotic pathway. In fact, expression of E2F1 is sufficient to induce this caspase induction.

So are caspases transcriptional targets of E2F1, or is the induction indirect? Northern blots revealed that caspase mRNA was increased by 5–15-fold (similar to the protein levels) when either E1A or E2F1 were expressed in IMR90 cells. Caspase mRNA levels also increase as cells enter S phase, which corresponds with the activity of E2F1 and the levels of cyclin A mRNA — a known E2F1 target. Analysis of caspase promoters provided further support that E2F could transcriptionally activate caspases, as several contain E2F1-binding sites, and chromatin immunoprecipitation experiments confirmed this — E2F1 precipitates from E1A-expressing cells that contained sequences from the caspase-7 promoter. The caspase-7 promoter was also able to drive transcription of the luciferase reporter gene — expression increased by almost 18-fold — when E2F1 was expressed.

But what is the physiological role of this caspase induction? It is not sufficient to induce apoptosis and, instead, seems to sensitize cells to apoptotic stimuli, such as serum withdrawal and adriamycin treatment. E2F1 is also known to activate cytochrome c release — a downstream event in the apoptotic pathway — via a p53-dependent pathway. The requirement for p53 in inducing apoptosis could be recapitulated, at least in part, by either introducing BAX — a proapoptotic protein that facilitates cytochrome c release — to TP53−/− BAX−/− cells expressing E2F, or by directly microinjecting cytochrome c into TP53−/− RB−/− cells. Under these conditions, caspase induction by E2F1 is able to enhance apoptosis, underscoring the cooperation of the p53-dependent and -independent pathways in inducing apoptosis.

So, the E1A oncogene coordinates division and death by using the same machinery — E2F — to initiate both processes. Whether other oncogenes operate in the same way remains to be determined.