**4. Summary and conclusions**

160 Steroids – Basic Science

phase and is controlled by mitotic cyclin/Cdk complexes, which are activated by removal of the inhibitory phosphates from Cdk1 by the Cdc25 phosphatases (eg. String in *Drosophila*) (Edgar and O'Farrell 1990). For cells to maintain their size, cell cycle progression must be accompanied by cell growth. However, during morphogenesis of the *Drosophila* abdominal epidermis from histoblasts, growth and division are uncoupled. The progenitor abdominal histoblasts are quiescent during the larval stages, but undergo rapid proliferation after pupation and eventually form the adult abdominal epidermis. Neither cell size nor division rate is constant for the developmentally regulated divisions that histoblast cells undergo during the larval and pupal stages. The onset of histoblast proliferation occurs 1–2 h after pupal formation (Ninov, Chiarelli, and Martin-Blanco 2007), which follows the ecdysone maximum at 0 h APF (Thummel 2001) and recent work has revealed that ecdysone is important for coupling growth and proliferation in abdominal histoblasts (Ninov, Manjon,

In contrast to the wing and eye epithelium, during larval stages histoblasts grow in a G2 arrested state prior to entering a proliferative stage during pupal metamorphosis (Hayashi 1996; Lawrence, Casal, and Struhl 1999, 1999). During larval stages the arrested histoblasts accumulate cellular mass in a process dependent on the insulin receptor/PI3K pathway and the transition to a proliferative state is initiated by ecdysone-dependent string/Cdc25 phosphatase transcription (Ninov, Manjon, and Martin-Blanco 2009). The latter can occur because the larval histoblasts have preaccumulated stores of the G1 cyclin, Cyclin E, which is sufficient to trigger S phase after mitosis. These cells show a progressive reduction of cell size as a consequence of the lack of a growth phase. After depletion of the stored Cyclin E, histoblasts proliferate more slowly and G1 is restored and cell proliferation again depends on growth factor signalling, requiring epidermal growth factor receptor (EGFR) signalling

during the G2/M transition and the insulin receptor/PI3K-pathway for growth.

Initiation of histoblast division by ecdysone/EcR occurs via transcriptional control of the cell cycle regulator String (Ninov, Manjon, and Martin-Blanco 2009). Previous work has shown that *string* overexpression triggers cell-cycle progression in embryonic and imaginal cells previously arrested in G2 (Edgar and O'Farrell 1990; Edgar, Lehman, and O'Farrell 1994; Milan, Campuzano, and Garcia-Bellido 1996), but not in G1-arrested cells (Kylsten and Saint 1997). Accordingly, the overexpression of String, but not Cyclin A, Cyclin B, or Cdk1, in histoblasts triggered their premature hyperproliferation in larval stages (Ninov, Manjon, and Martin-Blanco 2009). Although ecdysone is necessary to trigger histoblast proliferation (Ninov, Chiarelli, and Martin-Blanco 2007), upregulation of *string* transcription in larval stages bypasses the requirement for ecdysone pathway activity. As the block to histoblast proliferation following EcR knockdown with RNAi (Ninov, Chiarelli, and Martin-Blanco 2007) can be overcome by overexpression of *string*, which can still promote ectopic histoblast proliferation in the EcR loss of function cells (Ninov, Manjon, and Martin-Blanco 2009). As an indirect measure of *string* transcription a *string*-enhancer trap element was used, which revealed that EcR knockdown reduces *string* promoter activity. The authors also demonstrated reduced *string* mRNA levels by in situ hybridization. Further experiments are, however, required to determine whether these changes in *string* transcription are due to direct effects of EcR or mediated by another transcriptional regulator. Together this work revealed that the ecdysone pulse at the larval–pupal transition is required for the *string* transcription triggering histoblast proliferation at the onset of abdomen metamorphosis.

and Martin-Blanco 2009).

At the level of the whole animal, ecdysone controls larval growth and final body size through interactions with the insulin pathway (King-Jones and Thummel 2005; Shingleton 2005; Mirth and Riddiford 2007; Nijhout 2008). The insulin-signalling pathway acts in the prothoracic gland (PG) to regulate the release of ecdysone, therefore influencing the rate and duration of larval growth. For instance, PG overgrowth causes accelerated metamorphosis, which results in reduced adult size due to the rapid progression through the larval growth stage. Conversely, reducing growth of the PG results in longer larval growth periods and larger adults due to slower ecdysone release and delayed onset of pupariation. Correct timing of the critical peak in ecdysone is therefore essential for controlling larval growth and adult body size.

In the imaginal tissues and larval histoblasts ecdysone most likely regulates cell cycle genes indirectly by modulating upstream developmental signalling pathways. The effect of ecdysone on promoting SMW division in the eye may, in part, be mediated by Hedgehog (Hh) signalling (Brennan, Ashburner, and Moses 1998), and might coordinate this division with the G1 arrest in the furrow via the Dpp signal (Escudero and Freeman 2007; Firth and Baker 2005). In the wing imaginal disc, cell cycle progression requires EcR activity, which is associated with changes to the levels of *wingless* transcription. These changes in Wg may be mediated by the ecdysone-responsive transcription factor Crol (Mitchell et al. 2008) since EcR regulates cell cycle progression in a Crol dependent manner (Figure 7). Thus, by regulating the Wg pathway, which is known to control cell cycle in the wing (Johnston and Edgar 1998; Johnston and Sanders 2003; Herranz et al. 2008), the Crol transcription factor may provide a link between the ecdysone pulse and developmental cell cycle regulation in the wing (Figure 6; (Mitchell et al. 2008)). At the larval–pupal transition ecdysone activates *string* transcription in the histoblasts, triggering exit from G2 phase and histoblast proliferation. It will be of interest to determine whether these changes in *string* transcription are due to direct effects of EcR or, like the cell cycle changes occurring in imaginal tissues, are mediated by changes to developmental signalling.

Together the studies discussed here highlight the diverse mechanisms by which the ecdysone signal can impact on cell division in a range of tissues at different developmental time points. Further work is required to elucidate the molecular mechanisms underlying the ability of ecdysone to modify levels of the complex array of signals required for development.
