How do extra centrosomes initiate tumourigenesis in flies?

In WT 3rd instar neuroblasts there are always two centrioles, and, as cells enter mitosis, these are asymmetric, with one centriole organising more PCM than the other. In ~60% of neuroblasts that overexpress Sak, centrosomes are amplified, and there is no longer a single dominant centrosome. WT neuroblasts have many properties of stem cells, and they normally divide asymmetrically to regenerate a neuroblast and form a smaller ganglion mother cell. Approximately 10% of SakOE neuroblasts divide symmetrically, and neuroblast numbers are increased in SakOE brains. These brains can induce tumours when injected into the abdomen of WT hosts, indicating that centrosome amplification can induce tumour formation in flies (see Basto et al., 2008).

The idea that centrosome amplification can contribute to tumourigenesis was first proposed by Boveri almost one hundred years ago. He was aware that malignant cells often have an abnormal complement of chromosomes (a phenomenon termed “genetic instability”), and he proposed that extra centrosomes could drive genetic instability, because their presence invariably leads to chromosome missegregation in embryonic systems. Interest in Boveri’s hypothesis has recently been re-kindled with the realisation that centrosome amplification often occurs in cancer cells  and that the extent of centrosome amplification often correlates with the extent of genetic instability.

To test whether centrosome amplification inevitably leads to genetic instability and cancer, we generated stable Drosophila transgenic lines that over-express the centriole duplication protein Sak/Plk-4 (SakOE lines) (Basto et al., 2008). The over-expression of this protein drives centriole over-duplication, and ~60% of SakOE somatic cells have extra centrioles and centrosomes (see picture above). Surprisingly, SakOE flies are viable, fertile and morphologically normal, although they are significantly delayed in development. The cells with extra centrosomes initially form multipolar spindles, but they ultimately divide with a bipolar spindle, as most of the extra centrosomes become clustered into two spindle poles, and any centrosomes that are not clustered in this way become “inactivated” and lose their ability to organise MTs. Thus, although there is a small increase in the levels of aneuploidy in flies with extra centrosomes, centrosome amplification does not lead to large-scale genetic instability, and SakOE lines maintain a stable diploid genome over many generations.

Flies are often considered a poor model for human cancer because of their relatively short lifespan and because there is little cell division in adult flies, as most cell division stops when larvae pupate. Thus, adult flies almost never get cancer, and we failed to detect any obvious tumours in flies with extra centrosomes. The neoplastic growth of tissues does occur, however, in mutants such as lethal giant larvae (lgl)and discs large (dlg). This overgrowth usually takes place during a prolonged phase of larval development, which allows for an extended period of cell proliferation. To overcome the block to cell proliferation that occurs at pupation, one can use transplantation assays, in which larval tissues (either imaginal discs or brain) are transplanted into the abdomen of WT adult hosts. WT transplanted tissue can survive for several weeks in the host, but it does not form tumours. In contrast, the transplantation of tissue from various mutants leads to tissue over proliferation and the formation of metastatic tumours in the WT host.

To test whether SakOE larval brains form tumours when transplanted into WT hosts we expressed Tubulin-GFP (Tub-GFP) in SakOE flies so that we could follow the behaviour of the transplanted tissue. Whereas Tub-GFP brains behave normally after transplantation into WT hosts (n=90), 14% of Tub-GFP,SakOE brains form tumours under these conditions (n=104), and several form one or more metastases, far from the original site of injection (n = 5/15).

We were surprised that SakOE larval brains form tumours after transplantation as centrosome amplification does not generate large-scale genetic instability in flies. It has recently been shown, however, that brain tissue from several mutants in which the asymmetric divisions of the larval neuroblasts is perturbed can generate tumours in transplantation assays. It is thought that the defects in asymmetric division lead to an amplification of the neural stem cells (neuroblasts), and thereby to over-proliferation. We showed that neuroblasts with extra centrosomes have defects in asymmetric division: ~10% of SakOE neuroblasts divide symmetrically rather than asymmetrically, and there is an increase in neuroblast numbers in SakOE brains. Thus, we suspect that the tumourigenic potential of cells with extra centrosomes is due either to defects in the asymmetric division of the neuroblasts or to the small increase in genetic instability in these cells. We are currently trying to understand how centrosome amplification initiates tumourigenesis in flies.