Mice heterozygous for the retinoblastoma ( Rb ) gene develop pituitary and thyroid tumours on loss of the second allele. Inactivation of Trp53 in Rb-heterozygous mouse models has demonstrated some cooperative roles for loss of both tumour-suppressor genes, but not with respect to pituitary tumorigenesis. Kenneth Tsai and colleagues now report in the 24 December issue of Proceedings of the National Academy of Sciences that, surprisingly, loss of the tumour suppressor Arf — which activates p53 by inhibiting Mdm2 — accelerates pituitary tumorigenesis more than loss of Trp53 in Rb-heterozygous mice.

Tsai et al. found that loss of Arf accelerated pituitary tumorigenesis in Rb+/−Arf−/− mice and led to a significant decrease in their survival compared with control Rb+/− mice. The Rb+/−Arf−/− mice did not develop novel lesions that were characteristic of Rb+/−Trp53−/− mice, indicating that Arf loss is not equivalent to Trp53 loss in this model. The tumour cells from Rb+/−Arf−/− mice proliferated faster than those from the Trp53-null controls, but there was no difference in the rate of apoptosis.

Examination of the mice, 30 and 60 days after birth, showed that Arf-null mice developed atypical lesions earlier that were increased in number, larger and more aggressive. These results indicate that loss of Arf has different effects on an Rb-heterozygous phenotype than loss of Trp53. All early lesions had loss of heterozygosity (LOH) at Rb, showing that complete loss of Rb is still required for tumour formation in this model. The acceleration of tumorigenesis in the Rb+/−Arf−/− mice might be explained by Arf loss increasing proliferation of tumour cells after loss of heterozygosity at the Rb locus.

So, is Arf loss an obligatory event in pituitary tumorigenesis? It seems not, as in Rb heterozygotes that were also heterozygous for Arf, LOH for Arf occurred in only some of the tumours examined — so, unlike loss of Rb, Arf inactivation is not an essential step. However, Tsai et al. suggest that LOH of Arf might provide a selective advantage for early tumour cells.

The authors hypothesize that Arf might regulate a p53-independent mode of tumour suppression and suggest that this might be via regulation of p53-independent functions of the oncoprotein Mdm2.