Functional characterization of the alternative reading frame protein p14ARF

Abstract

A deeper understanding of the molecular events underlying tumor development is a prerequisite for the design of novel and efficient therapies. Inactivation of the p53 and retinoblastoma (Rb) tumor suppressor pathways appears as a common theme in most malignant human tumors. Most intriguingly, both p53 and Rb proteins are in part regulated through the CDKN2A (INK4a/ARF) locus on human chromosome 9p21, a region frequently lost in tumors. This region encodes two structurally and functionally distinct tumor suppressor proteins known as ARF (human p14ARF, mouse p19ARF), and p16INK4a. Exons 1 alpha, 2 and 3 encode p16INK4a, while exon 1 beta, spliced to exon 2 in an alternative reading frame encodes ARF. p16INK4a induces cell cycle arrest by inhibiting CDK4/6 whereas expression of ARF induces cell cycle arrest in part through inhibition of MDM2, a negative regulator of p53. This thesis is focused on the expression, localization and function of human p14ARF.

We discovered that p14ARF was overexpressed and localized to nucleoli, in human tumor cell lines deficient for p53 function as described in paper I. p14ARF had a similar intracellular localization as the major nucleolar phosphoprotein B23 (also known as nucleophosmin or NPM) during interphase, mitosis and in response to RNA polymerase I transcriptional inhibition that causes nucleolar dysfunction. B23 was identified as a bona fide p14ARF associated protein (paper II). The results indicated that B23 could be involved in efficient nucleolar localization and possibly stability of p14ARF protein. In paper III, the status of the p14ARF-MDM2-p53 pathway was investigated in a panel of Burkitt´s lymphoma (BL) cell lines. Loss of p14ARF (and p16INK4a) occurred in wildtype (wt) p53 containing BL lines only, whereas other BLs with wt p53 contained abundant levels of MDM2. Thus, inactivation of the p53 pathway was frequent in BL cell lines, presumably as a mechanism to escape or attenuate Myc induced p53-dependent apoptosis.

Next, we were interested in the role of p14ARF as a regulator of p53 activity in human fibroblasts after the activation of Myc or E2F-1 oncogenes (paper IV). Both Myc and E2F1 stabilized p53, along with phosphorylation on serine-15, induction of p21 and MDM2. Only E2F-1 markedly induced p14ARF. Both Myc and E2F-1 stabilized p53 in primary fibroblasts also after depletion of ARF. Caffeine blocked p53 accumulation after Myc or E2F-1 activation. Thus, the p53 response to activated oncogenes in normal human fibroblasts was not critically dependent on p14ARF. Interestingly, activation of Myc led to a strong accumulation of p16INK4a protein in primary human fibroblasts (paper V).

In summary, ARF is a nucleolar protein having some properties in common with previously characterized nucleolar proteins. But ARF is also a peculiar protein. Being both an inhibitor of MDM2 and of the ribosomal RNA processing machinery, ARF represents an interesting and almost unique link between p53 on one side, and the nucleolus on the other. The relative roles of ARF and p16INK4a in human tumor development remain enigmatic, but recent advances in the field indicate a predominant role of p16INK4a in protecting human cells from oncogenic transformation.