Cancer brings forward oviposition in the fly Drosophila melanogaster

Abstract Hosts often accelerate their reproductive effort in response to a parasitic infection, especially when their chances of future reproduction decrease with time from the onset of the infection. Because malignancies usually reduce survival, and hence potentially the fitness, it is expected that hosts with early cancer could have evolved to adjust their life‐history traits to maximize their immediate reproductive effort. Despite the potential importance of these plastic responses, little attention has been devoted to explore how cancers influence animal reproduction. Here, we use an experimental setup, a colony of genetically modified flies Drosophila melanogaster which develop colorectal cancer in the anterior gut, to show the role of cancer in altering life‐history traits. Specifically, we tested whether females adapt their reproductive strategy in response to harboring cancer. We found that flies with cancer reached the peak period of oviposition significantly earlier (i.e., 2 days) than healthy ones, while no difference in the length and extent of the fecundity peak was observed between the two groups of flies. Such compensatory responses to overcome the fitness‐limiting effect of cancer could explain the persistence of inherited cancer‐causing mutant alleles in the wild.

diseases and cancers, the impact of tumor development can be closely compared to that of infections by foreign organisms as neoplasia (i.e., new uncontrolled growth of cells that is not under physiologic control) broadly mimics their health as well as their fitness consequences (Vittecoq et al., 2015). Cancer, both solid and liquid tumors, can indeed be seen as an emerging species that behaves in a manner akin to parasites (Duesberg, Mandrioli, McCormack, & Nicholson, 2011). Although cancer is not transmitted, their evolution is governed by the same rules that apply to parasites: cancer cells depend on their hosts for sustenance, proliferating inside their bodies, exploiting them for energy and resources and thereby impairing their health and vigor. In the wild, due to interspecific interactions (especially predation and parasitism), death resulting from cancer per se is likely to occur prior to the end of the reproductive period (Vittecoq et al., 2015). Based on these ecological similarities with infectious diseases, we may predict that the responses evolved in the context of host-parasite interactions should also be relevant in the context of cancer. Although this hypothesis has been proposed (Boddy, Kokko, Breden, Wilkinson, & Aktipis, 2015;Crespi & Summers, 2006), to our knowledge, the adjustment of life-history traits in order to compensate for the negative fitness costs exerted by oncogenic processes has never been studied .
Using genetically modified fruit flies (Drosophila melanogaster) which develop malignant colorectal tumors, we explored the hypothesis of fecundity compensatory responses during the early stages of carcinogenesis. Knowing that APC-Ras flies display a reduced survival (~50% survival rate 4 weeks after cancer induction compared to ~70% survival rate for healthy flies; Martorell et al., 2014), we specifically tested whether females harboring tumors oviposit more and/or earlier than healthy ones before dying. Knowing that D. melanogaster has a fecundity peak early in life (Klepsatel et al., 2013), we expected three possible outcomes of female fly reproduction in response to cancer: (1) an earlier fecundity peak, (2) higher fecundity peak, or (3) a longer fecundity peak.

| Biological model
The genetically modified flies Drosophila melanogaster used in this study develop a colorectal cancer (CRC) of the anterior gut part. CRC starts with mutations in adenomatous polyposis coli (Apc) gene which encodes a large multidomain protein that plays an integral role in intercellular adhesion and in the wnt-signaling pathway (Nusse & Varmus, 2012). Deregulation of these processes causes unwanted cell growth which may induce tumor development (Fearnhead, Britton, & Bodmer, 2001;Logan & Nusse, 2004;Nusse & Varmus, 2012). The mutations in Apc gene are followed by the oncogenic activation of K-Ras gene (involved in the synthesis of protein component of the epidermal growth factor receptor signaling cascade), an event that correlates with the onset of malignancy (Karapetis et al., 2008;Ohnishil, Tomital, Monden, Yana, & Shimanol, 1997). For this research project, we used drosophila clones that combined the loss of Apc with the expression of the oncogenic form of Ras. These compound Apc-Ras clones, generated by a 1-hr heat shock at 37°C of 2-to 4-day-old drosophila males and females, expand as aggressive intestinal tumor-like overgrowths that reproduce many hallmarks of human CRC (Martorell et al., 2014;Wang et al., 2013). Previous experiments performed by us in routine in the laboratory indicated that 100% of the Apc-Ras clones experiencing a heat shock subsequently develop a colorectal cancer. Apc2N175K is a loss-of-function allele, ApcQ8 is a null allele, and UAS-RasV12 is a gain-of-function transgene. Flies were sex-separated at emergence, and MARCM clones were generated by a 1-hr heat shock at 37°C of 3-day-old Apc-Ras females and were marked by the progenitor cell marker escargot (esg) Gal4 line driving the expression of UAS-GFP.

| Experiment
Ten 3-day-old Apc-Ras virgin females or healthy virgin females were transferred into a small plastic cage (~540 cm 3 ) with 10 healthy males.
Two petri dishes containing fresh food (per liter of water: 10 g agar, 20 g sugar, and a drop of yeast was added when the medium is solidified)were fixed at the bottom of each cage. Petri dishes were replaced daily by fresh ones, and eggs were counted. Each treatment was replicated 36 times, and we followed the egg production over 2 weeks keeping the flies at 22°C (light/dark cycle 12 hr:12 hr). Dead flies (males and females) were also counted and removed daily to calculate the mean number of eggs laid by females. Males were replaced every 5-7 days by young males to keep sperm production constant.
Knowing that Apc-Ras and healthy flies are slightly genetically different, we controlled the effect of genetic background by repeating the same experiment, described above, without performing the heat shock (therefore not inducing cancer in the flies with Apc-Ras mutation). We repeated these two treatments 40 times each at 22°C and counted eggs laying over 7 days.

| Statistical analyses
Because our data violate the Gaussian assumption, we have adopted a nonparametric approach where (1) the average age at fecundity peak, (2) the average duration of the fecundity peak, and (3) the average clutch size (or number of eggs) at the fecundity peak were tested through a one-sided Mann-Whitney test. The duration of the fecundity peak was determined by identifying the last day when the slope of the linear regression between number of eggs laid and duration from the date of fecundity peak become significantly different from 0, that is, when the coefficient associated to the age from the linear regression is significant.
Considering the experiment without any heat shock (to control the effect of genetic background), there was no difference (W = 699.5, p-value = .8538) for the number of eggs laid between the Apc-Ras The duration of the fecundity peak has not been tested because only the first 8 days have been recorded, which is not enough to identify when the linear regression is not significant anymore.

| DISCUSSION
Responses displayed by animals in the face of cancer risks and/or malignant progression have received until now little attention (Vittecoq et al., 2015). The primary reasons this topic has been poorly investigated F I G U R E 1 Distribution of the different characteristics of the fecundity peak. (a) Date of fecundity peak (in number of days since the beginning of experiment). (b) The number of eggs at the fecundity peak (in number of eggs). (c) Duration of the fecundity peak (in number of days, see main text for a detailed explanation of this calculation). Only the date of fecundity peak is significantly different between cancerous (c) and noncancerous (NC) individuals (p-value = .025)