© 2001 Macmillan Magazines Ltd

Dogma would have it that both copies of atumour-suppressor gene must be inactivatedto promote tumour formation. But a newbreed of tumour suppressor is now emerging— for which inactivation of one allele isenough. In 1998, James Roberts, ChristopherKemp and their collaborators established thatCDKN1B, which encodes the cyclin-dependent kinase inhibitor Kip1, is haplo-insufficient for tumour suppression. Loss ofjust one copy of the Pten tumour suppressoris also sufficient to promote tumourprogression in a transgenic model of prostatecancer. Now, Kazushi Inoue and colleaguesdefine a new one-hit tumour suppressor inthe 15 November issue of Genes &Development. Most tumours downregulatethe p53 pathway by mutating or deletingboth TP53 alleles, expediting the disposal ofp53 by upregulating MDM2, or inactivating ARF, which inhibits MDM2. These proteinsare kept in check by a complex network ofcontrols: ARF, for example, is regulated byseveral transcriptional activators and

repressors, including the transcriptionalactivator DMP1.

Might DMP1 be a tumour suppressor inits own right? Inoue and colleaguespreviously reported that Dmp1–/– micedidn’t spontaneously develop tumours intheir first year of life, but now the mice areolder and we’re all a little wiser. In theirsecond year, Dmp1–/– mice spontaneouslydeveloped a variety of tumour types but,intriguingly, so did Dmp1+/– mice. This wasnot due to loss of heterozygosity orepigenetic silencing of the wild-type Dmp1allele because, in all tumours tested, thesecond allele was retained and mRNA andprotein were produced.

To explore whether Dmp1 mutanttumours select for loss of other genes in thep53 pathway, the authors crossed Dmp1mutant mice with Eµ-Myc mice, whichspontaneously develop Burkitt’slymphomas. Around half of these tumoursusually contain either p53 mutations or Arfmutations. Both Eµ-Myc/Dmp1+/– mice and

Eµ-Myc/Dmp–/– mice developed tumourswith a much shorter latency than Eµ-Myc/Dmp+/+ mice — 12 weeks ratherthan 6 months. Most of the tumours in theEµ-Myc/Dmp+/– mice produced detectablewild-type Dmp1 protein, supporting thenotion that losing just one Dmp1 allelepromotes tumour formation. Furthermore,only a small percentage of these mice (9%for Dmp1–/– mice; 14% for Dmp1+/– mice)sustained mutations in p53 or Arf.

Dmp1, then, is a bona fide tumoursuppressor and its downregulation, even by50%, reduces the selection pressure for loss ofp53 or Arf function. One of the central tenetsof cancer biology — Knudson’s two-hithypothesis — might just have sustainedanother hefty blow.

Cath Brooksbank

References and linksORIGINAL RESEARCH PAPER Inoue, K. et al. Dmp1 ishaplo-insufficient for tumor suppression and modifies thefrequencies of Arf and p53 mutations in Eµ-myc-inducedlymphomas. Genes Dev. 15, 2934–2939 (2001)FURTHER READING Fero, M. L. et al. The murine gene p27Kip1 is haplo-insufficientfor tumour suppression. Nature 396, 177–180 (1998) | Quon, K.& Berns, A. Haplo-insufficiency? Let me count the ways.Genes Dev. 15, 2917–2921 (2001) | Kwabi-Addo, B. et al.Haploinsufficiency of the Pten tumor suppressor genepromotes prostate cancer progression. Proc. Natl Acad. Sci.USA 98, 11563–11568 (2001)

Soft-tissue carcinomas (tumours ofmesodermal tissue) are — thankfully— rare. But when they do occur, theonly option is surgery. Survival ofindividuals with soft-tissue sarcomasis usually about 53 weeks, and treat-ment with conventional chemothera-py has so far been unsuccessful inpromoting survival. But now, encour-aging news from van Oosterom andcolleagues, reporting in the 27October issue of The Lancet, showsthat imatinib (STI-571 or Glivec) isboth safe and effective for the treat-ment of gastrointestinal stromaltumours (GISTs).

The United States Food and DrugAdministration approval of imatinibfor the treatment of chronic myel-ogenous leukaemia (CML) has creat-ed much excitement. As well asinhibiting BCR–ABL, it also blocksother oncogenic tyrosine kinases,including KIT, the receptor for stem-cell factor. KIT is overexpressed onmany — but not all — GISTs, sowould Glivec’s success in CML berepeated for this tumour? This Phase

I study aimed to establish a safe doselevel for treating GISTs.

van Oosterom and colleaguesenrolled 40 patients with sarcomas, 36of which were GISTs. The GISTpatients were all KIT positive. Patientswere given one of four dose levels:either 400 mg daily, 300 mg twicedaily, 400 mg twice daily or 500 mgtwice daily. Treatment was given untilthe tumours progressed, side effectsbecame too severe or the patientsrefused treatment. Otherwise,treatment continued for at least 1 year.

The main side effects during thefirst 8 weeks of treatment were skinrash, oedema, diarrhoea, nauseaand vomiting. More serious sideeffects, such as intratumoral bleed-ing, myelosuppression and neu-tropaenia occurred in a smallminority of patients, but mostimportantly, side effects were doselimiting at 500 mg imatinib twicedaily; at 400 mg imatinib twicedaily or less, side effects were man-ageable and diminished as treat-ment continued. So, 400 mg twice

Gut reaction

O N C O G E N E S

False-coloured 18fluorodeoxyglucose positron emission tomography scans of a patient whowas enrolled in the study, showing gastrointestinal stromal tumour status at the start of thestudy (left), after 8 days (middle) and after 4 weeks (right) of treatment with imatinib.Courtesy of Allan van Oosterom, Catholic University, Leuven, Belgium.

174 | DECEMBER 2001 | VOLUME 1 www.nature.com/reviews/cancer

H I G H L I G H T S

One-hit wonders?

T U M O U R S U P P R E S S O R S