Human tumours have a high degree of cellular and genetic heterogeneity, and the mechanisms by which distinct mutations in different cellular clones may cooperate to promote tumorigenesis are unclear. Xu and colleagues present evidence that a two-tier mechanism involving Jun N-terminal kinase (JNK) and Janus kinase (JAK)–signal transducer and activator of transcription (STAT) signalling enables interclonal cooperation between RasG12V and scribbled (scrib) mutations in Drosophila melanogaster.

Clones of genetically distinct cells can be created in D. melanogaster by induced mitotic recombination. Using this technique in the eye-antennal discs of D. melanogaster larvae, it has been shown that clones simultaneously expressing oncogenic RasG12V and harbouring loss-of-function mutations in the tumour suppressor scrib develop into large metastatic tumours (denoted RasG12Vscrib). Remarkably, Wu et al. now show that invasive tumours similarly develop in eye-antennal discs containing RasG12V and scrib mutations separately in adjacent clones (denoted RasG12V//scrib).

How do RasG12V and scrib mutations cooperate when they occur in different cells? Microarray analysis and real-time reverse transcription PCR revealed upregulation of the unpaired (Upd) family of cytokines, which activate JAK–STAT signalling, in RasG12V//scrib tumours. A dominant-negative form of the Upd receptor domeless, which mediates JAK–STAT activation, suppressed the growth and invasion of RasG12V//scrib tumours, and clones co-expressing RasG12V and Upd cytokines developed into metastatic tumours. Ras and JAK–STAT signalling therefore seem to have synergistic effects on tumorigenesis.

So how are Upd cytokines upregulated in RasG12V//scribtumours? Previous evidence shows that JNK upregulates Upd in wounds and that increased JNK signalling occurs in scrib clones. The authors therefore proposed that JNK activation in scrib cells induces Upd upregulation. In support of this, wing discs doubly mutant for scrib and the JNK kinase hemipterous (hep) had reduced activity of the JAK–STAT pathway. In addition, a dominant-negative form of the Jun kinase Basket (BskDN) suppressed RasG12V scrib tumours, but not RasG12V Upd tumours.

However, late-stage RasG12V//scrib tumours mostly contain RasG12V cells and few scrib cells. Therefore, JNK-dependent Upd upregulation in scrib cells cannot fully account for tumour development. Indeed, BskDN expression in RasG12V cells partially suppressed Upd upregulation and the growth of RasG12V//scrib tumours. This suggests that JNK-induced Upd expression occurs in RasG12V as well as scrib cells. Additionally, when wing discs were wounded in the anterior or posterior regions, JNK activity was observed across the whole disc. The authors therefore suggest that JNK activity in scrib cells can propagate JNK activation in adjacent RasG12V cells. In this way, JNK-dependent Upd upregulation in RasG12Vcells sustains tumour growth even when the original source of JNK activity, the scrib cells, is no longer present.

As tissue damage has been linked to human tumorigenesis, it is intriguing that wounding in wing discs also cooperated with RasG12V to promote overgrowth. Together, these results highlight the importance of cell interactions in tumour development, and it will be of interest to determine whether similar cooperative mechanisms contribute to human cancers.