Human cells are notoriously difficult to transform. Unlike rodent cells, which can be transformed with just a couple of oncogenes, human cells require several genetic lesions, including viral oncoproteins such as the SV40 large T and small t antigens, which disrupt both the p53 and RB tumour-suppressor pathways. Now, Gordon Peters and colleagues report in Cancer Cell that a new combination of just four genetic alterations can transform human diploid fibroblasts, and these leave the p53 pathway intact.

Culturing human cells results in a stress-induced, INK4A-dependent cell-cycle arrest, so the authors investigated the requirement for genetic lesions that induce transformation in Leiden cells, in which INK4A is specifically deleted. They had previously shown that immortalizing these cells with the telomerase subunit TERT and introducing the oncogene HRAS (LTR cells) allowed them to grow as anchorage-independent colonies. They now extended these observations by expressing, in LT cells, either MYC alone (LTM) or MYC and RAS (LTRM). These could proliferate in 0.2% agarose, so were also able to grow without anchorage. Western-blot analysis confirmed that members of the p53 pathway — ARF, p53 and WAF1 (also known as p21) — were expressed, so the cells could overcome the p53 checkpoint.

So, LTR, LTM and LTRM cells could all form anchorage-independent colonies, but were they fully transformed? Only the LTRM cells had a transformed phenotype — they were smaller and more rounded — and resulted in tumours on subcutaneous injection into nude mice. However, as only 5/16 innoculations gave rise to tumours, and the latency was relatively long — 59–98 days — another alteration might have occurred.

The phenotype of the tumour cells was further analysed by re-plating them into tissue culture. They had reduced adhesion and produced autocrine growth factors — they could proliferate in low serum. The tumours also had high levels of p53 and its targets, MDM2 and WAF1.

As the tumour cells had maintained a functional p53 pathway, might they also have avoided the aneuploidy that frequently characterizes tumour cells? A combination of multiplex fluorescence in situ hybridization and comparative genomic hybridization revealed that although the cells were normally diploid, two changes — on chromosomes 18 and 20 — were frequently observed.

Although these results leave us with some unanswered questions, they do provide a more physiological starting point for examining the mechanisms and consequences of transformation, and highlight some differences between the processes that occur in mouse and human cells.